InfoVis:Wiki - User contributions [en]

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:52:36Z

UE-InfoVis0809 9825992: vertical dashed line

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

==== Details and Examples ====
The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
For the quantitative data set the visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

For events, the instances are mapped to horizontal bars, which extend from their start date to their end date. For very short time interval (e.g., birth of Barth) the bar made wider, so that the event is visible.

==== Techniques ====
The application uses the following visualization techniques:
* Line charts for time-series [Few, 2004a]
* LifeLines [Plaisant et al., 1998]
* Focus and context [Leung and Apperley, 1994]

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.
It is possible to observe weekly patterns in consumption of beer and coffee.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple.
We want to avoid situations, where the users might get lost in a specific view / selection and not find their way back.

Still, we provide the following interaction modes:
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** The visualizations in all tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) buttons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** The user can zoom in and zoom out by clicking on the (+) and (-) buttons.
** Panning is possible by clicking on the < and > buttons.
** By clicking on a LifeLine bar, the user can move the focus of the visualisation on that time span.
* Details on demand
** If the user moves the mouse pointer over a line, a small popup will show the value at this point of time (Figure 1).
** Moving the mouse over a LifeLine bar, the exact start and end date will be shown in a popup window.
* Switching future scenarios: By selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios (Figure 4).
** A vertical dashed line separates past data from estimated future values.

=== Discussion ===
Our tool supports users by presenting relevant data at a glance. Charts showing important quantitative data are shown in alignment a view of the patient's social environment, which helps to identify events and periods with significant impact on the patients health.

The screen setup is kept simple, thus focusing the user's attention to the relevant data.

While focusing on data relevant for physicians most charts do also have a meaning for patients. Especially charts rising significantly over the ''100%''-limit emphasize the need for a diet change.

==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, Visual Analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:40:37Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

==== Details and Examples ====
The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
For the quantitative data set the visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

For events, the instances are mapped to horizontal bars, which extend from their start date to their end date. For very short time interval (e.g., birth of Barth) the bar made wider, so that the event is visible.

==== Techniques ====
The application uses the following visualization techniques:
* Line charts for time-series [Few, 2004a]
* LifeLines [Plaisant et al., 1998]
* Focus and context [Leung and Apperley, 1994]

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.
It is possible to observe weekly patterns in consumption of beer and coffee.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple.
We want to avoid situations, where the users might get lost in a specific view / selection and not find their way back.

Still, we provide the following interaction modes:
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** The visualizations in all tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) buttons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** The user can zoom in and zoom out by clicking on the (+) and (-) buttons.
** Panning is possible by clicking on the < and > buttons.
** By clicking on a LifeLine bar, the user can move the focus of the visualisation on that time span.
* Details on demand
** If the user moves the mouse pointer over a line, a small popup will show the value at this point of time (Figure 1).
** Moving the mouse over a LifeLine bar, the exact start and end date will be shown in a popup window.
* Switching future scenarios: By selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios (Figure 4).

=== Discussion ===
Our tool supports users by presenting relevant data at a glance. Charts showing important quantitative data are shown in alignment a view of the patient's social environment, which helps to identify events and periods with significant impact on the patients health.

The screen setup is kept simple, thus focusing the user's attention to the relevant data.

While focusing on data relevant for physicians most charts do also have a meaning for patients. Especially charts rising significantly over the ''100%''-limit emphasize the need for a diet change.

==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, Visual Analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:36:42Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

==== Details and Examples ====
The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
For the quantitative data set the visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

For events, the instances are mapped to horizontal bars, which extend from their start date to their end date. For very short time interval (e.g., birth of Barth) the bar made wider, so that the event is visible.

==== Techniques ====
The application uses the following visualization techniques:
* Line charts for time-series [Few, 2004a]
* LifeLines [Plaisant et al., 1998]
* Focus and context [Leung and Apperley, 1994]

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.
It is possible to observe weekly patterns in consumption of beer and coffee.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple.
We want to avoid situations, where the users might get lost in a specific view / selection and not find their way back.

Still, we provide the following interaction modes:
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** The visualizations in all tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) buttons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** The user can zoom in and zoom out by clicking on the (+) and (-) buttons.
** Panning is possible by clicking on the < and > buttons.
** By clicking on a LifeLine bar, the user can move the focus of the visualisation on that time span.
* Details on demand
** If the user moves the mouse pointer over a line, a small popup will show the value at this point of time (Figure 1).
** Moving the mouse over a LifeLine bar, the exact start and end date will be shown in a popup window.
* Switching future scenarios: By selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios (Figure 4).

=== Discussion ===
Our tool supports users by presenting relevant data at a glance. Charts showing important quantitative data are shown in alignment a view of the patient's social environment, which helps to identify events and periods with significant impact on the patients health.

==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, Visual Analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:29:53Z

UE-InfoVis0809 9825992: patterns in figure2

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

==== Details and Examples ====
The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
For the quantitative data set the visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

For events, the instances are mapped to horizontal bars, which extend from their start date to their end date. For very short time interval (e.g., birth of Barth) the bar made wider, so that the event is visible.

==== Techniques ====
The application uses the following visualization techniques:
* Line charts for time-series [Few, 2004a]
* LifeLines [Plaisant et al., 1998]
* Focus and context [Leung and Apperley, 1994]

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.
It is possible to observe weekly patterns in consumption of beer and coffee.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple.
We want to avoid situations, where the users might get lost in a specific view / selection and not find their way back.

Still, we provide the following interaction modes:
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** The visualizations in all tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) buttons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** The user can zoom in and zoom out by clicking on the (+) and (-) buttons.
** Panning is possible by clicking on the < and > buttons.
** By clicking on a LifeLine bar, the user can move the focus of the visualisation on that time span.
* Details on demand
** If the user moves the mouse pointer over a line, a small popup will show the value at this point of time (Figure 1).
** Moving the mouse over a LifeLine bar, the exact start and end date will be shown in a popup window.
* Switching future scenarios: By selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios (Figure 4).

=== Discussion ===
Our tool supports users by presenting relevant data at a glance. Important charts are shown within the context of the patients environment which helps to identify events and periods with significant impact on the patients health.

==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, Visual Analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:27:28Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

==== Details and Examples ====
The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
For the quantitative data set the visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

For events, the instances are mapped to horizontal bars, which extend from their start date to their end date. For very short time interval (e.g., birth of Barth) the bar made wider, so that the event is visible.

==== Techniques ====
The application uses the following visualization techniques:
* Line charts for time-series [Few, 2004a]
* LifeLines [Plaisant et al., 1998]
* Focus and context [Leung and Apperley, 1994]

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple.
We want to avoid situations, where the users might get lost in a specific view / selection and not find their way back.

Still, we provide the following interaction modes:
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** The visualizations in all tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) buttons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** The user can zoom in and zoom out by clicking on the (+) and (-) buttons.
** Panning is possible by clicking on the < and > buttons.
** By clicking on a LifeLine bar, the user can move the focus of the visualisation on that time span.
* Details on demand
** If the user moves the mouse pointer over a line, a small popup will show the value at this point of time (Figure 1).
** Moving the mouse over a LifeLine bar, the exact start and end date will be shown in a popup window.
* Switching future scenarios: By selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios (Figure 4).

=== Discussion ===
Our tool supports users by presenting relevant data at a glance. Important charts are shown within the context of the patients environment which helps to identify events and periods with significant impact on the patients health.

==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, Visual Analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:23:47Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

==== Details and Examples ====
The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
For the quantitative data set the visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

For events, the instances are mapped to horizontal bars, which extend from their start date to their end date. For very short time interval (e.g., birth of Barth) the bar made wider, so that the event is visible.

==== Techniques ====
The application uses the following visualization techniques:
* Line charts for time-series [Few, 2004a]
* LifeLines [Plaisant et al., 1998]
* Focus and context [Leung and Apperley, 1994]

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple.
We want to avoid situations, where the users might get lost in a specific view / selection and not find their way back.

Still, we provide the following interaction modes:
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** The visualizations in all tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) buttons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** The user can zoom in and zoom out by clicking on the (+) and (-) buttons.
** Panning is possible by clicking on the < and > buttons.
** By clicking on a LifeLine bar, the user can move the focus of the visualisation on that time span.
* Details on demand
** If the user moves the mouse pointer over a line, a small popup will show the value at this point of time.
** Moving the mouse over a LifeLine bar, the exact start and end date will be shown in a popup window.
* Switching future scenarios: By selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios.

=== Discussion ===
Our tool supports users by presenting relevant data at a glance. Important charts are shown within the context of the patients environment which helps to identify events and periods with significant impact on the patients health.

==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, visual analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.
* [Pownser and Tufte, 1994] Seth M. Powsner and Edward R. Tufte. Graphical Summary of Patient Status. ''The Lancet'', 344(8919):386-389, August 1994.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T23:01:31Z

UE-InfoVis0809 9825992: Type of Visualization

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====

Kosara and Miksch [2002] describe three types of applications for Information Visualization in medicine: visualizing measurements, displaying incidents, and planing the future. This application requires the visualization of quantitative values and events. Therefore it combines charts with a LifeLines inspired view of events [Plaisant et al., 1998].

The quantitative data set is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

==== Visual Mapping ====
The visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

LifeLines [Plaisant et al., 1998]

==== Techniques used ====
* Focus and context [Leung and Apperley, 1994]
* Mouse Actions: we employ actions like clicking and mouse-over.
* Time-Series [Few, 2004a]

* Beschreibung der verwendeten Techniken / angewandten Prinzipien
** inkl. Quellenangaben (siehe z.B. VO Skriptum/Foliensatz)

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple, to avoid situations where the users might get lost in a specific view / selection and not finding their way back.

Still, we provide the following interaction modes
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** These tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) icons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** By clicking on the LifeLine element, the user can move the focus of the visualisation on that time span
* Details on demand
* Switching future scenarios: with selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios.

=== Discussion ===
* Wie werden die BenutzerInnen in ihren Aufgaben unterstützt?
* Welche Besonderheiten gibt es?
==== Advantages and drawbacks ====
* Our tool is simple, and clear in presentation, with only showing one chart per screen.
* We focus on showing only the most relevant visualisations of the data, while skipping e.g. summary charts for a specific period, or displaying numerical data. This reduces the needed interaction and number of available system states significantly.
* However, showing more charts per screen would allow the users to visualize several aspects of the data at once, and maybe get some insights that might be more difficult to achieve otherwise. However, clarity and the need for interaction would clearly raise.

=== Future Work ===
As future work on this prototype, we should carry out user studies to back up or change some of the design decisions, namely:
* A comparison of the currently employed "tabs" to change the visualized data with displaying all charts on one screen, with the possibility to collapse/expand each of them, should be carried out.
* Additionally, alternative user-interface elements than the tabs, e.g. radio-buttons, might be tested
* An evaluation of the "focus and context" method, possibly comparing it to other zooming techniques.

Further, it might be worth to provide the users with annotation tools, to support documentation of their own insights as well as facilitating communication and discussion on a specific patient with colleagues.

Finally, visual analytic methods should be utilised to help the users find patterns in the data.

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.
* [Pownser and Tufte, 1994] Seth M. Powsner and Edward R. Tufte. Graphical Summary of Patient Status. ''The Lancet'', 344(8919):386-389, August 1994.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T22:40:27Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the lump sugar.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| lump sugar || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====
The data sets is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two can be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

Events -> LifeLines [Plaisant et al., 1998]

==== Visual Mapping ====
The visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

LifeLines [Plaisant et al., 1998]

==== Techniques used ====
* Focus and context [Leung and Apperley, 1994]
* Mouse Actions: we employ actions like clicking and mouse-over.
* Time-Series [Few, 2004a]

* Beschreibung der verwendeten Techniken / angewandten Prinzipien
** inkl. Quellenangaben (siehe z.B. VO Skriptum/Foliensatz)

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the focus and context technique.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple, to avoid situations where the users might get lost in a specific view / selection and not finding their way back.

Still, we provide the following interaction modes
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** These tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) icons. Zooming uses the "focus and context" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** By clicking on the LifeLine element, the user can move the focus of the visualisation on that time span
* Details on demand
* Switching future scenarios: with selecting one of the options from the drop-down box on the top-right, the user can select one of the future scenarios.

=== Discussion ===

* Wie werden die BenutzerInnen in ihren Aufgaben unterstützt?
* Welche Besonderheiten gibt es?
* Was sind die Vor- und Nachteile der Technik?

=== Future Work ===

* Erweiterungs- bzw. Verbesserungsmöglichkeiten

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.
* [Pownser and Tufte, 1994] Seth M. Powsner and Edward R. Tufte. Graphical Summary of Patient Status. ''The Lancet'', 344(8919):386-389, August 1994.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T22:26:28Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe4.html Beschreibung der Aufgabe 4]

=== Gegebene Daten ===

'''Homer Simpson's Trinkverhalten in Abhängigkeit von seinen Lebensumständen''' 

...Visualisierung von Homer's Lebensabschnitten bzw. Ereignissen mit Einfluss auf sein
Trinkverhalten (zB.: Kindheit, Pubertät, Arbeitslosigkeit, Beziehungen, Hochzeit, Geburt
der Kinder, Liebeskummer, Alltag, etc.) von seiner Geburt bis Jetzt + mögliche
Zukunftsszenarien (mind. 3).
 
*Die Menge folgender Getränke soll für die jeweiligen Lebensumstände ablesbar sein
(ml oder Liter - je nachdem - pro Tag, Monat, Jahr (z.B.: Fokus+Kontext Methoden):
a) Wasser 
b) Milch 
c) Fruchtsaft 
d) Cola 
e) Kaffee (Würfelzucker?) 
f) Bier
(vereinfacht angenommen, Homer trinkt ausschließlich diese Getränke)
 
*Die folgenden Werte sollen abhängig von den konsumierten Getränken ablesbar sein:
1) g oder kg konsumierter Zucker (aus Getränken) + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 50g; enthaltener Zucker: 10g/100 ml Cola; 10g/100 ml Fruchtsaft; 3g/Würfelzucker). 
2) mg konsumiertes Coffein + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 600mg; enthaltenes Coffein: 10 mg/100 ml Cola; 80 mg/100 ml Kaffee). 
3) g konsumierter Alkohol + empfohlene Maximaldosis pro Tag, Monat, Jahr
(empfohlene Maximaldosis/Tag: 20g; enthaltener Alkohol: 3,6 g/100ml Bier)
 
*Die Daten sollen zur medizinischen/psychologischen Analyse visualisiert werden. 
*Die bisher erlernten Design-Prinzipien sollen umgesetzt werden (z.B.: Optimierung der Data-ink ratio). 
*Die Mockups sollten zumindest 1) Homer's Leben im Überblick 2) und eine Detailansicht wiedergeben. 
*Alle nicht angeführten Daten können frei erfunden werden.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== General description and background ==

=== Field of application ===
The visualization will be used by experts for medical and psychological analysis. The visualization shall provide information of a single patient, and give details about their consumption of several different drinks (water, milk, juice, coke, coffee, and beer), and the thus resulting consumption of sugar, caffeine and alcohol. It shall be possible to visualize both a time-line overview of these consumptions, as well as detailed views for specific periods in time. Additionally, certain periods of life, such as childhood or puberty, and specific events and happenings such as unemployment, wedding(s), or birth of the children, shall be indicated.

=== Analysis of the data set ===

The application displays ''abstract'', ''multi-dimensional'', ''time-oriented'' information:

* '''Drinks''': Consumption of six different drinks (water, milk, juice, coke, coffee, and beer) and sugar cubes by day.
** These dimensions have a ''continuous'' data type.
* '''Ingredients''': The amount of dangerous substances (sugar, caffeine, and alcohol) contained in the drinks and in the sugar cubes.
** These dimensions have a ''continuous'' data type.
** These variables are ''derived'' from the drinks and sugar cubes.
* '''Events''': Certain periods of life and specific events (e.g., childhood, unemployment).
** This dimension is of a ''nominal'' data type; they can be distinguished, but they cannot be ordered, as they might be overlapping.

The following table gives a detailed overview on each datum:

{|cellpadding="2" border="1"
! Dimension !! Data type !! Unit
|-
| water || continuous || ml
|-
| milk || continuous || ml
|-
| juice || continuous || ml
|-
| coke || continuous || ml
|-
| coffee || continuous || ml
|-
| beer || continuous || ml
|-
| sugar cube || discrete || pieces
|-
| sugar || continuous || g
|-
| caffeine || continuous || g
|-
| alcohol || continuous || g
|-
| events || nominal || - |
|}

==== Temporal Dimension ====

We can characterize the temporal dimension as follows [Aigner, 2006, pp. 13-17]:

* The application's time domain is ''branching'', because it allows the description of alternative scenarios.
* The time scale is ''discrete'' and days are the smallest granularity used.
* Drinks and ingredients are ''totally ordered'', whereas events are ''partially ordered'', because they might be overlapping.
* The temporal primitives used for drinks and ingredients are ''instants'', whereas ''intervals'' are used for events.
* The application uses the ''Gregorian calendar'' and different temporal granularities like months and year will be made available to the users.

The ingredients of the drinks, and their recommended maximum dose per day, are as follows:

{| border="1"
! !! Sugar !! Caffeine !! Alcohol
|-
! Water
| || ||
|-
! Milk
| || ||
|-
! Juice (100ml)
| align="right" | 10g || ||
|-
! Coffee (100ml)
| || align="right" | 80mg ||
|-
! Lump sugar (per piece)
| align="right" | 3g || ||
|-
! Coke (100ml)
| align="right" | 10g || align="right" | 10mg ||
|-
! Beer (100ml)
| align="right" | || || align="right" | 3.6g
|-
|-cellpadding="2"

! Daily dose
| align="right" | 50g || align="right" | 600mg || align="right" | 20g
|}

For the events and periods in life, the data would look like:

{| border="1"
! Start !! End !! Type !! remarks
|-
| 13.4.1965 || 12.4.1972 || Childhood ||
|-
| 7.7.1977 || 7.7.1977 || Wedding ||
|-
| 25.7.1980 || 25.7.1980 || Birth || Bart
|-
| 1.8.1981 || 31.10.1982 || Unemployment ||
|-
| 12.10.1982 || 12.10.1982 || Birth || Lisa
|-
|}

An exemplary data set for the consumption would look as follows:

{| border="1"
! Date !! Water !! Milk !! Juice !! Coffee !! Lump sugar !! Coke !! Beer
|-
| 7.7. 1977 || 0.25 || 0.5 || 0 || 0.75 || 5 || 1 || 5.5
|-
| 25.7. 1980 || 0 || 0 || 0.25 || 1.25 || 12 || 0 || 0
|-
| 1.1. 1993 || 1.0 || 0 || 0.25 || 0.25 || 2 || 0.33 || 1.5
|-
|}

=== Analysis of the target group ===
The intended target group of the visualization are physicians, psychologists and nutritionist. Thanks to their education they can interpret of the data and thus do primary need hints what different values imply but require guidance to navigate through the data and conceive it.

On the other hand it cannot be assumed that the users have technical knowledge of interactive systems or are willing to acquire it, therefore the visualization must be intuitive and as far as possible self descriptive.

Furthermore possible limitations by the target hardware must be taken into account. This could also limit the possibilities of interaction (e.g. screen-size, no keyboard or no mouse but a touch-pad).

=== Goals of the visualization ===
The visualisation shall allow medical doctors to get a fast overview on excess consumption of certain drinks and the thus resulting consumption of substances, potentially to be able to link it with certain symptoms and diseases Homer might suffer from.

For psychologists, it might be more relevant to link the consumption to events, to find patterns in what circumstances Homer tends to have an excess consumption of the certain drinks. It might be especially interesting to spot excess consumption that cannot be linked with a certain event, for example too much beer consumption that isn't in relation to a celebration or a known phase of depression, like unemployment.

Generally, we assume the doctors to use these tools mainly during a physical examination, or maybe a ward round; we thus expect the doctors having a limited amount of time available for inspecting the data, thus they should be able to get a fast overview and fast zooming into specific time periods. Therefore, we want to provide for a simple navigation, and rather focus on displaying data in the context of the whole life span, rather than giving (too) detailed data or even numbers for a specific date.

== Concept ==

=== Visualization ===
==== Type of Visualization ====
The data sets is composed of multiple instances of several measures taken periodically, i.e. a time series, which is best visualised as a line chart [Few, 2004].
We have several different aspects to visualize
* Consumption of drinks
* Consumption of substances/ingredients
* Individual consumption of caffeine, sugar and alcohol, indicating the source of the substance

The first two cam be visualized with normal line charts, while for the last one, we opted for a stacked line chart.

Events -> LifeLines [Plaisant et al., 1998]

==== Visual Mapping ====
The visual mapping is the same in all our visualisation variants, and rather simple: the time series is mapped to the horizontal axis [Few, 2004], while the vertical axis maps the values at the specific point in time. The scale of the vertical axis either measured in litres consumption per day, or as a percentage of the recommended daily dose. The recommended dose, which is at 100%, is specifically highlighted, to allow for easily spotting over-consumption.

LifeLines [Plaisant et al., 1998]

* Beschreibung der verwendeten Techniken / angewandten Prinzipien
** inkl. Quellenangaben (siehe z.B. VO Skriptum/Foliensatz)

=== Mockup(s) ===
The imaginary data-set and the charts were produced using Open Office Calc, while the final mockups were created with a drawing-program.

[[Image:Ue08-simpson-overview.png]] 
Figure 1: Overview and standard visualisation, showing the consumption of drinks during the whole life of the patient.

[[Image:Ue08-simpson-detail.png]] 
Figure 2: Detailed view of a specific period of time, using the context and zoom technique.

[[Image:Ue08-simpson-ingredients.png]] 
Figure 3: Overview, showing the consumption of substances.

[[Image:Ue08-simpson-caffeine.png]] 
Figure 4: Overview, showing the sources of caffeine consumed.

=== Interaction ===
As the application is meant for medical doctors and psychologists, we kept the interface simple, to avoid situations where the users might get lost in a specific view / selection and not finding their way back.

Still, we provide the following interaction modes
* Switching between different data to visualise: with the tabs on the top-right part of the tool, just above the time-line. This allows the user to select either the consumption of drinks, consumption of substances, or detailed consumption and sources for each of the substances caffeine, sugar and alcohol.
** These tabs are connected with each other: If the user zooms in a tab and then changes to another tab, the same timespan will be focussed.
* Zooming into a specific period of life: this can be initiated by clicking on a specific location of the time-line, and by using the (+) and (-) icons. Zooming uses the "context and zoom" technique to always view the selected area in context with the data of the other stages of life of the patient [Leung and Apperley, 1994].
** click on LifeLine element to move focus on that time span
* Details on demand
* Switching future scenarios

=== Discussion ===

* Wie werden die BenutzerInnen in ihren Aufgaben unterstützt?
* Welche Besonderheiten gibt es?
* Was sind die Vor- und Nachteile der Technik?

=== Future Work ===

* Erweiterungs- bzw. Verbesserungsmöglichkeiten

== References ==
* [Aigner, 2006] Wolfgang Aigner. Visualization of Time and Time-Oriented Information: Challenges and Conceptual Design. Ph.D. thesis, Vienna University of Technology, 2006.
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.
* [Kosara and Miksch, 2002] Robert Kosara and Silvia Miksch. Visualization Methods for Data Analysis and Planning in Medical Applications. ''International Journal of Medical Informatics'', 68(1-3):141-153, December 2002.
* [Leung and Apperley, 1994] Y.K. Leung and M.D. Apperley. A Review and Taxonomy of Distortion-Oriented Presentation Techniques. ''ACM Transactions on Computer-Human Interaction (TOCHI)'', 1(2):126-160, June 1994.
* [Plaisant et al., 1998] Catherine Plaisant, Richard Mushlin, Aaron Snyder, Jia Li, Dan Heller, and Ben Shneiderman. LifeLines: Using Visualization to Enhance Navigation and Analysis of Patient Records. Technical Report HCIL-98-08, University of Maryland, 1998.
* [Pownser and Tufte, 1994] Seth M. Powsner and Edward R. Tufte. Graphical Summary of Patient Status. ''The Lancet'', 344(8919):386-389, August 1994.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T20:43:59Z

UE-InfoVis0809 9825992: insert mockups

File:Ue08-simpson-overview.png

2009-01-07T20:41:07Z

UE-InfoVis0809 9825992: Homer Simpson's consumption of drinks over his whole lifespan

== Summary ==
Homer Simpson's consumption of drinks over his whole lifespan
== Copyright status ==

== Source ==
created with Inkscape, OpenOffice and GIMP

File:Ue08-simpson-ingredients.png

2009-01-07T20:39:53Z

UE-InfoVis0809 9825992: Homer Simpson's consumption of caffeine, sugar, and alcohol over his whole lifespan

== Summary ==
Homer Simpson's consumption of caffeine, sugar, and alcohol over his whole lifespan
== Copyright status ==

== Source ==
created with Inkscape, OpenOffice and GIMP

File:Ue08-simpson-detail.png

2009-01-07T20:38:43Z

UE-InfoVis0809 9825992: Homer Simpson's consumption of drinks over his whole lifespan with focus on 4 weeks

== Summary ==
Homer Simpson's consumption of drinks over his whole lifespan with focus on 4 weeks
== Copyright status ==

== Source ==
created with Inkscape, OpenOffice and GIMP

File:Ue08-simpson-caffeine.png

2009-01-07T20:37:19Z

UE-InfoVis0809 9825992: Homer Simpson's caffeine level over his whole lifespan

== Summary ==
Homer Simpson's caffeine level over his whole lifespan
== Copyright status ==

== Source ==
created with Inkscape, OpenOffice and GIMP

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T16:44:25Z

UE-InfoVis0809 9825992: more references

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T13:02:02Z

UE-InfoVis0809 9825992: more references

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T12:54:52Z

UE-InfoVis0809 9825992: references

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T12:48:12Z

UE-InfoVis0809 9825992: structure

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T12:47:40Z

UE-InfoVis0809 9825992: structure

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T12:15:30Z

UE-InfoVis0809 9825992: citation

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T10:59:41Z

UE-InfoVis0809 9825992: formating to put emphasis

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-07T10:57:01Z

UE-InfoVis0809 9825992:

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 4

2009-01-06T22:53:30Z

UE-InfoVis0809 9825992: data set

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-29T17:33:22Z

UE-InfoVis0809 9825992: Improvements

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [p. 187].
** We noticed that the pies are drawn in (at least) two subtly different variations of red. The pies for "firefighter", "clergy", "athlete", and "actor" appear to be brighter. We found this disturbing, as we could not find out if this difference encodes any information.
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

[[Image:callings.png]]

== Improvements ==

* We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
** We used blue color for the bars, removed the bar borders, and set a 1-to-1.5 ratio of bar to intervening white space.
** We also colored the bar for "teachers" blue, because there is no obvious reason for highlighting "teachers".
** We didn't highlight the bars for "firefighter", "clergy", "athlete", and "actor", because we were not sure, if the difference in the original graphic encode any information.
* On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
* To ease comparison and look-up of values, we used thin grid lines.
* We didn't include the values in numbers, because we believe the purpose of this graphic is to communicate ranking and relation between values, but not exact values.
* Categorical labels were written in mixed case to improve legibility.
* We changed the title from "Callings" to "Prestigious Professions", which is more meaningful.
* We reduced the font size of the reference to the data source.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 2

2008-12-15T09:36:58Z

UE-InfoVis0809 9825992: work in tutor's feedback

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe2.html Beschreibung der Aufgabe 2]

=== Zu beurteilende Tabelle ===
[[Image:ImageRRH.JPG]]

== Critical Review ==
We reviewed the given table based on the table design guidelines by Few [2004] and found the following problems:

=== Delineating Columns and Rows ===
A grid is used to delineating columns and rows, which should be avoided.

=== Arranging Data ===
The table has multiple columns called "Rank", which contain the same data.
We don't know if this is by coincidence, or redundancy.

The columns have different widths which groups (and separates) the data. This seems not to be on purpose.

The data rows are not sorted by any column. Readers will spend some time trying to understand, why the rows are printed in this order. Further, if readers wants to compare values, they can do it more efficiently, if rows are sorted by rank.

=== Formatting Text ===
All columns containing numbers are centered.
Some numbers (e.g., 4.3) only have one digit right of the decimal point. Consequently these numbers are narrower than the other numbers in the column and are not well aligned.

Some of the columns (Rank) have a smaller font size.
We compared the asterisk symbols and found out that three different font families are used.

The table contains footnotes.

== Improved Table ==
[[Image:ImageRRH-1.png]]

== Improvements ==

=== Delineating Columns and Rows ===
We removed the grid and slightly increased line height.

Rulers are used to seperate the header from the table-body.

Additional whitespace was added between the columns "All Ss" and "Hard News" to enforce the bond between "Hard News" and "Soft News". The same was done for "HN-SN", "t" and "α" resp.

=== Arranging Data ===
By inserting whitespace we created three column groups for "All Ss", "Hard News", and "Soft News" to structure the data.
Each of these groups contain a column for the value and a column for the rank.
As said above, it is unclear whether the rank is redundant (and should kept only once) or describes three different rankings. We assumed the latter, thus kept all columns.

Instead of "M1" and "M2" we used the variables "HN" and "SN". These akronyms are explained in the table notes.

We reordered the data rows by rank. Since the three rankings could be as well different it would be also possible to order by the first column.

To ease lookup of "α", we created an additional column.

=== Formatting Text ===
We used only one font family and one font size
We applied the same number format and precision for all values to improve alignment.

Columns containing numbers are aligned to the right.

The resulting table can be browsed easily and looks good.
Even if it's size would be downscaled by ~15%, the content would be comprehensible.

== Links ==
* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==
* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-09T18:00:02Z

UE-InfoVis0809 9825992: extend improvements

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [Few, 2004, p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [Few, 2004, p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [Few, 2004, p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [Few, 2004, p. 187].
** We noticed that the pies are drawn in (at least) two subtly different variations of red. The pies for "firefighter", "clergy", "athlete", and "actor" appear to be brighter. We found this disturbing, as we could not find out if this difference encodes any information.
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

[[Image:callings.png]]

== Improvements ==

* We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
** We used blue color for the bars, removed the bar borders, and set a 1-to-1.5 ratio of bar to intervening white space.
** We colored the bar for "teachers" orange.
** We didn't highlight the bars for "firefighter", "clergy", "athlete", and "actor", because we were not sure, if the difference in the original graphic encode any information.
* On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
* To ease comparison and look-up of values, we used thin grid lines.
* We didn't include the values in numbers, because we believe the purpose of this graphic is to communicate ranking and relation between values, but not exact values.
* Categorical labels were written in mixed case to improve legibility.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-09T17:48:40Z

UE-InfoVis0809 9825992:

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [Few, 2004, p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [Few, 2004, p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [Few, 2004, p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [Few, 2004, p. 187].
** We noticed that the pies are drawn in (at least) two subtly different variations of red. The pies for "firefighter", "clergy", "athlete", "actor" appear to be brighter. We found this disturbing, as we could not find out if this difference encodes any information.
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

[[Image:callings.png]]

== Improvements ==

* We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
** We used blue color for the bars, removed the bar borders, and set a 1-to-1.5 ratio of bar to intervening white space.
** We colored the bar for "teachers" orange.
* On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
* To ease comparison and look-up of values, we used thin grid lines.
* Categorical labels were written in mixed case to improve legibility.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

File:Callings .png

2008-12-09T17:46:51Z

UE-InfoVis0809 9825992: New page: == Summary == == Copyright status == == Source ==

== Summary ==

== Copyright status ==

== Source ==

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-09T13:43:50Z

UE-InfoVis0809 9825992: variations of red

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [Few, 2004, p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [Few, 2004, p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [Few, 2004, p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [Few, 2004, p. 187].
** We noticed that the pies are drawn in (at least) two subtly different variations of red. The pies for "firefighter", "clergy", "athlete", "actor" appear to be brighter. We found this disturbing, as we could not find out if this difference encodes any information.
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

[[Image:callings.png]]

[[Image:callings2.png]]

== Improvements ==

* We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
** We used blue color for the bars, removed the bar borders, and set a 1-to-1.5 ratio of bar to intervening white space.
** We colored the bar for "teachers" orange.
* On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
* To ease comparison and look-up of values, we used thin grid lines.
* Categorical labels were written in mixed case to improve legibility.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-08T21:58:07Z

UE-InfoVis0809 9825992: some improvements

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [Few, 2004, p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [Few, 2004, p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [Few, 2004, p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [Few, 2004, p. 187].
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

[[Image:callings.png]]

[[Image:callings2.png]]

== Improvements ==

* We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
** We used blue color for the bars, removed the bar borders, and set a 1-to-1.5 ratio of bar to intervening white space.
** We colored the bar for "teachers" orange.
* On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
* To ease comparison and look-up of values, we used thin grid lines.
* Categorical labels were written in mixed case to improve legibility.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

File:Callings2.png

2008-12-08T21:54:25Z

UE-InfoVis0809 9825992: Callings - Proportion of respondents who attribute 'very great' prestige to professionals

== Summary ==
Callings - Proportion of respondents who attribute 'very great' prestige to professionals
== Copyright status ==

== Source ==
Graphic by Alexander Rind; Data by: The Harris Poll, July 2008 created with OpenOffice 2.4

File:Callings.png

2008-12-07T18:00:32Z

UE-InfoVis0809 9825992:

== Summary ==
Callings - Proportion of respondents who attribute 'very great' prestige to professionals

== Copyright status ==

== Source ==
Graphic by Alexander Rind; Data by: The Harris Poll, July 2008

created with OpenOffice 2.4

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-07T17:59:16Z

UE-InfoVis0809 9825992: Improved Graphic

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [Few, 2004, p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [Few, 2004, p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [Few, 2004, p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [Few, 2004, p. 187].
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

[[Image:callings.png]]

== Improvements ==

* We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
** We used blue color for the bars, turned off borders and set a 1-to-1.5 ratio of bar to intervening white space.
** We colored the bar for "teachers" orange.
* On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
* To ease comparison and look-up of values, we used thin grid lines.
* Categorical labels were written in mixed case to improve legibility.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

File:Callings.png

2008-12-07T17:54:32Z

UE-InfoVis0809 9825992: Callings - Proportion of respondents who attribute 'very great' prestige to professionals

== Summary ==
Callings - Proportion of respondents who attribute 'very great' prestige to professionals

== Copyright status ==

== Source ==
Graphic by Alexander Rind; Data by: The Harris Poll, July 2008

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-07T17:49:23Z

UE-InfoVis0809 9825992: Improvements

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe3.html Beschreibung der Aufgabe 3]

=== Zu beurteilende Grafik ===
[[Image:Nytimes-callings.jpg]] 
Proportion of respondents who attribute 'very great' prestige to professionals

== Critical Review ==

Based on the design principles recorded by Few [2004], we discovered the following problems in the graphic shown above:

* Data objects with '''3-D depth''' are used. These objects are hard to read. Furthermore, the graphical components added for 3-D depth do not convey any extra information (bad data ink ratio) [Few, 2004, p. 171].
* Through the use of '''pie charts''', quantitative information is encoded by area. These types of graphs communicate poorly, because the human visual perception cannot accurately estimate the area of a 2-D shapes. In 3-D this is even harder [Few, 2004, p. 60].
* The numeric values are printed next to the categorical labels. This decreases the '''data-ink ratio'''. If we take into account the two points mentioned above, we can argue that the pie charts are non-data ink and only categorical labels and values are data ink.
* Except for "teachers" all pies are drawn in an '''intense color''' (highly saturated red). This draws a lot of attention to these pies, especially because red signals danger in most western cultures [Few, 2004, p. 171]. However, we think that the graphic's purpose is to highlight the proportion for "teachers". In that case the pie for "teachers" should be more intense [Few, 2004, p. 187].
* The categorical labels are printed in '''capitals'''. In this form text is less legible.

== Improved Graphic ==

== Improvements ==

# We decided to use horizontal bars, because they are well suited for displaying ranking relationships [Few, 2004, p. 73, 182].
## We used blue color for the bars, turned off borders and set a 1-to-1.5 ratio of bar to intervening white space.
## We colored the bar for "teachers" orange.
# On the categorical axis we switched off tick marks, because they are not needed. On the quantitative axis we put tick marks every 10%.
# To ease comparison and look-up of values, we used thin grid lines.
# Categorical labels were written in mixed case to improve legibility.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]
* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]
* [[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 3

2008-12-07T16:58:49Z

UE-InfoVis0809 9825992: Critical Review

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 1 - Small Multiples

2008-11-26T17:17:35Z

UE-InfoVis0809 9825992: fixed typos

[[Image:Tufte 2001 small-multiple la-pollution.png|thumb|300px|Pollution [Tufte, 2001]]]

[[Image:Tufte 2001 small-multiple cars.png|thumb|300px|Car repairs [Tufte, 2001]]]

[[Image:Powsner_1994_medical.png|thumb|300px|Medical record [Pownser and Tufte, 1994]]]

{{Definition|In the small multiples method, multiple postage-stamp size representations, positioned within eyespan, show different parts of the data in a constant design.}}

== Principles ==
Small multiples, a representation principle introduced by Edward R. Tufte [1990, 2001], make it easy to compare different products, different parameters, or how parameters change over time.
A constant visual structure allows the human reader to focus on changes in the data without being disturbed by changes in the scales, axes, etc.
The small size makes it possible to put many representation within eyespan.
Spreading small multiples across several pages would eliminate these effects.

Small multiples use a constant visual structure to show different parts of the data.
In contrast, representations based on [[Multiple Views]] use different visual structures to show the same data in all views.

{{Quotation|Illustrations of postage-stamp size are indexed by category or a label, sequenced over time like the frames of a movie, or ordered by a quantitative variable not used in the single image itself.|[Tufte, 1990]}}

== Examples ==
The position of a small multiples in the two dimensional representation space can be mapped to data dimension.
Consequently, small multiples are well suited for representing multivariate data.
For example, in the pollution visualization (top right) the individual images are indexed by time and pollutant. Within each graphic pollution levels are shown on a two dimensional spatial grid.
Thus, a total of 28,800 data values are shown, except for those hidden by peaks [Tufte, 2001].

The second example (middle right) shows the frequency-of-repair records for cars built from 1976 to 1981.
Edward Tufte [2001] describes it as "a particularly ingenious mix of table and graphic, portraying a complex set of comparisons".

The third example is the visualization of a patient's medical record.
It consists of small scatterplots to show medical status and drug dosages over time.
These values are normalised to allow for easy comparison.
Furthermore related parameters are grouped: e.g, Insulin is located directly below Glucose [Pownser and Tufte, 1994].

== References ==

*[Pownser and Tufte, 1994] Seth M. Powsner and Edward R. Tufte. Graphical Summary of Patient Status. ''The Lancet'', 344(8919):386-389, August 1994.
*[Tufte, 2001] Edward R. Tufte. ''The Visual Display of Quantitative Information'', Second Edition. Graphics Press, Cheshire, CT, 2001.
*[Tufte, 1990] Edward R. Tufte. [[Tufte, E.R.: Envisioning Information, Graphics Press, 1990|''Envisioning Information'']]. Graphics Press, Cheshire, CT, 1990.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 2

2008-11-21T00:30:29Z

UE-InfoVis0809 9825992: make sentences

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe2.html Beschreibung der Aufgabe 2]

=== Zu beurteilende Tabelle ===

[[Image:ImageRRH.JPG]]

== Critical Review ==

We reviewed the given table based on the table design guidelines by Few [2004] and found the following problems:

=== Delineating Columns and Rows ===
A grid is used to delineating columns and rows, which should be avoided.

=== Arranging Data ===
The table has multiple columns called "rank", which have the same data.
We don't know, if this is by coincidence, or redundance.

The data rows are not sorted by any column.

=== Formatting Text ===

All columns containing numbers have centered.
Some numbers (e.g., 4.3) only have one digit right of the decimal point. Consequently these numbers are narrower than the other numbers in the column and are not well aligned.

Some of the columns (rank) have a smaller font size.
We compared the asterisk symbols and found out that three different font families are used.

The table contains footnotes.

== Improved Table ==

[[Image:ImageRRH-1.png]]

== Improvements ==

=== Delineating Columns and Rows ===

We removed the grid and slightly increased line height.
Rulers are used in the header.

=== Arranging Data ===

We created three column groups for "All Ss", "Hard News", and "Soft News".
Each of these groups contain a column for the value and a column for the rank.
Instead of "M1" and "M2" we used the variables "HN" and "SN".

We reordered the data rows by rank.

To avoid footnotes we created an additional column.

=== Formatting Text ===

We used only one font family and one font size
We applied the same number format and precision for all values to improve alignment.

The resulting table can be browsed easily and looks good.
Even if it's size would be downscaled by ~15%, the content would be comprehensible.

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 2

2008-11-20T22:54:55Z

UE-InfoVis0809 9825992: reference

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe2.html Beschreibung der Aufgabe 2]

=== Zu beurteilende Tabelle ===

[[Image:ImageRRH.JPG]]

== Critical Review ==

* grid
* multiple cols with redundant data(?)
* numbers
** centered alignment
** mixed decimals
* no structural elements
* ambiguous Column headers
* data-rows in unnatural order (none at all?)
* footnotes
* different fonts and fontsizes

== Improved Table ==

[[Image:ImageRRH-1.png]]

== Improvements ==

* no grid
* only one font, one size
* structural elements for header
** seperated header
** grouped columns (values and rank)
* reordered (rank)
* inlined alpha
* same formatting and precision in every column --> improves alignment
* slightly increased line height

overall size of the table could be downscaled ~15% (?)

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

* [Few, 2004] Stephen Few. Show Me the Numbers - Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, CA, 2004.

Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06 - Aufgabe 2

2008-11-18T19:02:28Z

UE-InfoVis0809 9825992: page structure

== Aufgabenstellung ==
[http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/infovis_ue_aufgabe2.html Beschreibung der Aufgabe 2]

=== Zu beurteilende Tabelle ===

[[Image:ImageRRH.JPG]]

== Critical Review ==

== Improved Table ==

== Improvements ==

== Links ==

* [[Teaching:TUW_-_UE_InfoVis_WS_2008/09|InfoVis:Wiki UE Homepage]]

* [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws08/ UE InfoVis]

*[[Teaching:TUW - UE InfoVis WS 2008/09 - Gruppe 06|Gruppe 06]]

== References ==

File:Tufte 2001 small-multiple cars.png

2008-11-07T19:26:13Z

UE-InfoVis0809 9825992:

== Summary ==
Small multiples to visualize frequency-of-repair for cars
== Copyright status ==

== Source ==
Edward R. Tufte. ''The Visual Display of Quantitative Information'', Second Edition. Graphics Press, Cheshire, CT, 2001, page 174.

File:Tufte 2001 small-multiple la-pollution.png

2008-11-07T19:25:13Z

UE-InfoVis0809 9825992:

== Summary ==
Small multiples to visualize pollution in southern California
== Copyright status ==

== Source ==
Edward R. Tufte. The Visual Display of Quantitative Information, Second Edition. Graphics Press, Cheshire, CT, 2001, page 42.

File:Smith 2001 treemap usenet.jpg

2008-11-07T19:21:16Z