InfoVis:Wiki - User contributions [en]

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Prototype -Documentation

2006-12-18T12:48:09Z

UE-InfoVis0607 0026116: /* Documentation */

===Documentation===
Projectname:

authors:
*Taskdescription
Our task was to implement a prototype of a conversation landscape visualisation.

Conversation Landscapes are a method to visualize archives of Chats. In this 2D modell, chat user are arranged along the x axis and the y axis represent time. The horizontal lines are postings, a longer line represents a longer posting.

The viewer can easily read out the time between postings as periods of silence. Convesation Landscapes show the interaction patterns of the conversation at a glance, flurries of animated discussion are as easily readable as logins and logouts.

*Implementation
*Screenshot
*Used Toolkits bzw. Libraries
*Links to further documentations
[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Prototype -Documentation

2006-12-18T12:46:23Z

UE-InfoVis0607 0026116: /* Documentation */

===Documentation===
Projectname:

authors:
*Taskdescription
Our task was to implement a prototype of a conversation landscape visualisation.

Conversation Landscapes are an interface to visualize the archive of Chat Cirles. In this 2D modell, chat user are arranged along the x axis and the y axis represent time. The horizontal lines are postings, a longer line represents a longer posting.

The viewer can easily read out the time between postings as periods of silence. Convesation Landscapes show the interaction patterns of the conversation at a glance, flurries of animated discussion are as easily readable as logins and logouts.

*Implementation
*Screenshot
*Used Toolkits bzw. Libraries
*Links to further documentations
[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Prototype -Documentation

2006-12-18T12:44:00Z

UE-InfoVis0607 0026116: /* Documentation */

===Documentation===
Projectname:

authors:
*Taskdescription
*Implementation
*Screenshot
*Used Toolkits bzw. Libraries
Links to further documentations

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Prototype -Documentation

2006-12-18T12:43:33Z

UE-InfoVis0607 0026116: documentation

===Documentation===
Projectname:
authors:
*Taskdescription
*Implementation
*Screenshot
*Used Toolkits bzw. Libraries
Links to further documentations

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T17:04:36Z

UE-InfoVis0607 0026116: /* Chat Circles */

==Conversation Landscape ==
=== Chat Cirles Description ===
*Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===
*chat interface
every person who is participating in the chat is represented with a coloured circle, the color is unique for each user.
the messages of this user appears in this circle.
the frequence of messages, who activ a user is, is shown this way, the circle gets bigger and brighter with every message and smaller and starts to fade within periodes of silence, but it is still visuable.
the different topics are shown as groups of circles, because the user cann only "hear" others in a range around the own position.

*archive interface
in the chat archive interface of chat circles, again color is used to identify users.

=== Interaction ===

====User-Interaction====
the user can choose the color of his/her circle.
the user can move the own circle around.

====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T12:04:38Z

UE-InfoVis0607 0026116: /* Visual Mapping */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===
*chat interface
every person who is participating in the chat is represented with a coloured circle, the color is unique for each user.
the messages of this user appears in this circle.
the frequence of messages, who activ a user is, is shown this way, the circle gets bigger and brighter with every message and smaller and starts to fade within periodes of silence, but it is still visuable.
the different topics are shown as groups of circles, because the user cann only "hear" others in a range around the own position.

*archive interface
in the chat archive interface of chat circles, again color is used to identify users.

=== Interaction ===

====User-Interaction====
the user can choose the color of his/her circle.
the user can move the own circle around.

====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:49:46Z

UE-InfoVis0607 0026116: /* User-Interaction */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===
every person who is participating in the chat is represented with a coloured circle.
the messages of this user appears in this circle.
the frequence of messages, who activ a user is, is shown this way, the circle gets bigger and brighter with every message and smaller and starts to fade within periodes of silence, but it is still visuable.
the different topics are shown as groups of circles, because the user cann only "hear" others in a range around the own position.

=== Interaction ===

====User-Interaction====
the user can choose the color of his/her circle.
the user can move the own circle around.

====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:48:45Z

UE-InfoVis0607 0026116: /* Visual Mapping */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===
every person who is participating in the chat is represented with a coloured circle.
the messages of this user appears in this circle.
the frequence of messages, who activ a user is, is shown this way, the circle gets bigger and brighter with every message and smaller and starts to fade within periodes of silence, but it is still visuable.
the different topics are shown as groups of circles, because the user cann only "hear" others in a range around the own position.

=== Interaction ===

====User-Interaction====

====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:46:11Z

UE-InfoVis0607 0026116: /* Visual Mapping */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===
every person who is participating in the chat is represented with a coloured circle.
the messages of this user appears in this circle.
the frequence of messages, who activ a user is, is shown this way, the circle gets bigger and brighter with every message and smaller and starts to fade within periodes of silence, but it is still visuable.

=== Interaction ===

====User-Interaction====

====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:38:02Z

UE-InfoVis0607 0026116: /* Interaction */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===

=== Interaction ===

====User-Interaction====

====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:37:16Z

UE-InfoVis0607 0026116: /* Description */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===

=== Interaction ===
====User-Interaction====
====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:36:54Z

UE-InfoVis0607 0026116: /* Description */

==Chat Circles ==
=== Description ===
* Chat Circles is a Technique for vizualising Chats.
Mainly to get an better overview of the social interactions taking place in the chat.

=== Visual Mapping ===

=== Interaction ===
====User-Interaction====
====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T11:03:27Z

UE-InfoVis0607 0026116: /* Description */

==Chat Circles ==
=== Description ===
* Zusammenfassung und Beschreibung der gewählten InfoVis Technik
o inkl. Bilder und Quellenangaben
=== Visual Mapping ===

=== Interaction ===
====User-Interaction====
====Use of Interactions====
* Navigation
* Zooming
* Highlighting
* Dynamic Querying
* Selection
* Brushing

==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T10:45:22Z

UE-InfoVis0607 0026116: /* Loom */

==Chat Circles ==
=== Description ===
* Zusammenfassung und Beschreibung der gewählten InfoVis Technik
o inkl. Bilder und Quellenangaben
* Visual Mapping
o Wie werden die verschiedenen Datendimensionen in visuelle Attribute umgesetzt?
* Interaktionsmöglichkeiten
o Wie erfolgt die Interaktion der BenutzerInnen mit dem System?
o Welche Interaktion dient welchem Zweck?
+ Navigation, Zooming, Highlighting, Dynamic Querying, Selection, Brushing, ...
==Loom==
=== Description ===

== References ==

:[Donath et al., 1999] Donath, J., Karahalios, K., and Viegas, F.: Visualizing Conversation. In Proceedings of the 32nd Annual Hawaii International Conference on System Sciences (HICSS) - Volume 2, page 2023. IEEE Computer Society, 1999.http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/papers/ConversationLandscapes_VisualizeConv.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique

2006-11-27T10:40:01Z

UE-InfoVis0607 0026116:

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3

2006-11-27T10:37:15Z

UE-InfoVis0607 0026116: uebersicht

=== Aufgabe 3 ===
*'''[[Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 3 - Technique|Technique]]'''

Teaching:TUW - UE InfoVis WS 2006/07 - Aufgabe 3 - Technikauswahl

2006-11-27T10:33:44Z

UE-InfoVis0607 0026116:

{|
|-
|'''Gruppe: '''
|'''Gewählte Technik:'''
|-
|01
|2 SpiraClock
|-
|02
|4 Tile Maps oder Arc Diagram
|-
|03
|13 Conversation Landscape
|-
|08
|3 Perspective Wall oder 6 ThemeRiver
|-
|}

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T21:56:09Z

UE-InfoVis0607 0026116: /* Data Density */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor Graphic==

===Principles of Graphical Integrity===
* '''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is about 57% and the effect in the graphic about 55%, which results in a Lie Factor of nearly 1. The Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005], so the impression of the two graphs seems to be an optical illusion. Conclusion: Don't use graphform which tend to give the viewers wrong impressions of size differences.
* Because of the 3D-effect in the graphic the object in the background would be much more bigger in the foreground. Unfortunately the 3D-effect doesn't really come out at first sight and so the viewer compares the object directly.
* The numbers are not readable and the lines between the numbers are very irritating and confusing.
* As well is the colour gradient. The gradient goes in the smaler "3D Hurricane" from blue to black and in the bigger one reverse.
* a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
* A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
* The white lines forming squares (left) do not contain additional information. They only try to create the 3D-effect.
* The background color has no additional information.
* It's not necessary to write "CAT. 4-5 Hurricanes" twice.
* In this case you could remove the numbers for the regions, because they are unreadable.

'''Missing ink'''
* The ink doesn't change as the data changes.

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
* The "3D Hurricanes" take a lot of space. Its better to replace them with smaller representations of the data, e.g. bars.
* Divide the regions in separate values.

===Chartjunk===
* The white grid and the shadow in the background is used to imitate a 3D-effect. The 3D-effect is not needed in this graphic and so the grid should be junked.
* In the text a reason for increased hurricanes is mentioned, but the shown data displays no evidence for that, because there is no related data.
* The text "bigger, faster, stronger, more" is not necessary. The graphic shows (or should show) that there are more hurricanes than in the last years.

=== Design ===
* The graphic isn't very comparative. The smaller "3D Hurricane" appears to be just farther behind in time and position.
* Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
* Although the data is about hurricanes it doesn't suggest to use the shape of a hurricane in the graphic. It's better to use a horizontal graphic about 50% wider than tall.

* The type is not very clear or modest.
* The word "category" is not spelt out.

* The data is wrong, the sum of the hurricanes (171) as shown in the left graphic does not equal the summed up hurricans from the regions (151) shown in thetext of the same graphic.
* Some text in the graphic is written in uppercase only letters. That's difficult to read.

==Improved Graphic==

We made two improved graphics, the difference between them is that the first focuses on the comparability of the values in the different world regions and the second on the comparability of the entire number of hurricanes.

=== First improved graphic: Focus on the different world regions ===

[[Image:Verbessert3.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparison of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

=== Second improved graphic: Focus on the whole numbers of hurricanes ===

[[Image:Verbessert2.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparison of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

==Description of Improvements==

* Use of a 2-dimensional graph instead of a pseudo-3d graph
* Dividing the regions into seperate values.
* Use of different colors in the comparision between the entire number of hurricanes.
* Text in upper-and lowercase. We didn't use a serif font because on a computer screen a font without serifs is easier to read.

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

[Few, 2004] Stephen Few, Elegance Through Simplicity, intelligent enterprise, Oct. 16, 2004, http://www.intelligententerprise.com/showArticle.jhtml;jsessionid=N2ATDQWY5VYKSQSNDBGCKHSCJUMEKJVN?articleID=49400920

[Tufte, 1999] Edward R. Tufte: The Visual Display of Quantitative Information, Stanford University, Jan. 26, 1999, http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T21:53:00Z

UE-InfoVis0607 0026116: /* Data Density */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor Graphic==

===Principles of Graphical Integrity===
* '''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is ca.57% and the effect in the graphic about 55%, which results in a Lie Factor of nearly 1.The Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005], so my impression of the two graphs seems to be an optical illusion. Conclusion: Don't use graphform which tend to give the viewers wrong impressions of size differences.
* Because of the 3d-effect in the graphic the object in the background would be much more bigger in the foreground. Unfortunately the 3d-effect doesn't really come out at first sight and so the viewer compares the object directly.
* The numbers are not readable and the lines between the numbers are very irritating and confusing.
* As well is the colour gradient. The gradient goes in the smaler "3D Hurricane" from blue to black and in the bigger one reverse.
* a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
* A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
* The white lines forming squares (left) do not contain additional information.
* The background color has no additional information.
* It's not necessary to write "CAT. 4-5 Hurricanes" twice.
* In this case you could remove the numbers for the regions, because they are unreadable.

'''Missing ink'''
* The ink doesn't change as the data changes.

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
* The "3D Hurricanes" take a lot of space. Its better to replace them with smaler representations of the data, e.g. bars.
* divide the regions in seperate values.

===Chartjunk===
* The white grid and the shadow in the background is used to imitate a 3d-effect. The 3d-effect is not needed in this graphic and so the grid should be junked.
* In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
* The text "bigger, faster, stronger, more" is not necessary. The graphic shows (or should show) that there are more Hurricanes than in the last years.

=== Design ===
* The graphic isn't very comparative. The smaller "3D Hurricane" appears to be just farther behind in time and position.
* Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
* Although the data is about hurricans it doesn't suggets to use the shape of a hurrican in the graphic. It's better to use a horizontal graphic about 50% wider than tall.

* The type is not very clear or modest.
* The word "category" is not spelt out.

* The data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(151) shown in the same graphic.
* Some text in the graphic is written in uppercase only letters. That's difficult to read.

==Improved Graphic==

We made two improved graphics, the difference between them is that the first focuses on the comparability of the values in the different world regions and the second on the comparability of the entire number of hurricanes.

=== First improved graphic: Focus on the different world regions ===

[[Image:Verbessert3.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

=== Second improved graphic: Focus on the whole numbers of hurricanes ===

[[Image:Verbessert2.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

==Description of Improvements==

* Use of a 2-dimensional graph instead of a pseudo-3d graph
* Dividing the regions into seperate values.
* Use of different colors in the comparision between the entire number of hurricanes.
* Text in upper-and lowercase. We didn't use a serif font because on a computer screen a font without serifs is easier to read.

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

[Few, 2004] Stephen Few, Elegance Through Simplicity, intelligent enterprise, Oct. 16, 2004, http://www.intelligententerprise.com/showArticle.jhtml;jsessionid=N2ATDQWY5VYKSQSNDBGCKHSCJUMEKJVN?articleID=49400920

[Tufte, 1999] Edward R. Tufte: The Visual Display of Quantitative Information, Stanford University, Jan. 26, 1999, http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T21:51:53Z

UE-InfoVis0607 0026116: /* Design */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor Graphic==

===Principles of Graphical Integrity===
* '''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is ca.57% and the effect in the graphic about 55%, which results in a Lie Factor of nearly 1.The Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005], so my impression of the two graphs seems to be an optical illusion. Conclusion: Don't use graphform which tend to give the viewers wrong impressions of size differences.
* Because of the 3d-effect in the graphic the object in the background would be much more bigger in the foreground. Unfortunately the 3d-effect doesn't really come out at first sight and so the viewer compares the object directly.
* The numbers are not readable and the lines between the numbers are very irritating and confusing.
* As well is the colour gradient. The gradient goes in the smaler "3D Hurricane" from blue to black and in the bigger one reverse.
* a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
* A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
* The white lines forming squares (left) do not contain additional information.
* The background color has no additional information.
* It's not necessary to write "CAT. 4-5 Hurricanes" twice.
* In this case you could remove the numbers for the regions, because they are unreadable.

'''Missing ink'''
* The ink doesn't change as the data changes.

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
* The "3D Hurricanes" take a lot of space. Its better to replace them with smaler representations of the data, e.g. bars.
* divide the regions in seperate

===Chartjunk===
* The white grid and the shadow in the background is used to imitate a 3d-effect. The 3d-effect is not needed in this graphic and so the grid should be junked.
* In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
* The text "bigger, faster, stronger, more" is not necessary. The graphic shows (or should show) that there are more Hurricanes than in the last years.

=== Design ===
* The graphic isn't very comparative. The smaller "3D Hurricane" appears to be just farther behind in time and position.
* Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
* Although the data is about hurricans it doesn't suggets to use the shape of a hurrican in the graphic. It's better to use a horizontal graphic about 50% wider than tall.

* The type is not very clear or modest.
* The word "category" is not spelt out.

* The data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(151) shown in the same graphic.
* Some text in the graphic is written in uppercase only letters. That's difficult to read.

==Improved Graphic==

We made two improved graphics, the difference between them is that the first focuses on the comparability of the values in the different world regions and the second on the comparability of the entire number of hurricanes.

=== First improved graphic: Focus on the different world regions ===

[[Image:Verbessert3.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

=== Second improved graphic: Focus on the whole numbers of hurricanes ===

[[Image:Verbessert2.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

==Description of Improvements==

* Use of a 2-dimensional graph instead of a pseudo-3d graph
* Dividing the regions into seperate values.
* Use of different colors in the comparision between the entire number of hurricanes.
* Text in upper-and lowercase. We didn't use a serif font because on a computer screen a font without serifs is easier to read.

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

[Few, 2004] Stephen Few, Elegance Through Simplicity, intelligent enterprise, Oct. 16, 2004, http://www.intelligententerprise.com/showArticle.jhtml;jsessionid=N2ATDQWY5VYKSQSNDBGCKHSCJUMEKJVN?articleID=49400920

[Tufte, 1999] Edward R. Tufte: The Visual Display of Quantitative Information, Stanford University, Jan. 26, 1999, http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T21:50:56Z

UE-InfoVis0607 0026116: /* Data-Ink Ratio */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor Graphic==

===Principles of Graphical Integrity===
* '''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is ca.57% and the effect in the graphic about 55%, which results in a Lie Factor of nearly 1.The Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005], so my impression of the two graphs seems to be an optical illusion. Conclusion: Don't use graphform which tend to give the viewers wrong impressions of size differences.
* Because of the 3d-effect in the graphic the object in the background would be much more bigger in the foreground. Unfortunately the 3d-effect doesn't really come out at first sight and so the viewer compares the object directly.
* The numbers are not readable and the lines between the numbers are very irritating and confusing.
* As well is the colour gradient. The gradient goes in the smaler "3D Hurricane" from blue to black and in the bigger one reverse.
* a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
* A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
* The white lines forming squares (left) do not contain additional information.
* The background color has no additional information.
* It's not necessary to write "CAT. 4-5 Hurricanes" twice.
* In this case you could remove the numbers for the regions, because they are unreadable.

'''Missing ink'''
* The ink doesn't change as the data changes.

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
* The "3D Hurricanes" take a lot of space. Its better to replace them with smaler representations of the data, e.g. bars.
* divide the regions in seperate

===Chartjunk===
* The white grid and the shadow in the background is used to imitate a 3d-effect. The 3d-effect is not needed in this graphic and so the grid should be junked.
* In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
* The text "bigger, faster, stronger, more" is not necessary. The graphic shows (or should show) that there are more Hurricanes than in the last years.

=== Design ===
* The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
* Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
* Although the data is about hurricans it doesn't suggets to use the shape of a hurrican in the graphic. It's better to use a horizontal graphic about 50% wider than tall.

* The type is not very clear or modest.
* The word "category" is not spelt out.

* The data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(151) shown in the same graphic.
* Some text in the graphic is written in uppercase only letters. That's difficult to read.

==Improved Graphic==

We made two improved graphics, the difference between them is that the first focuses on the comparability of the values in the different world regions and the second on the comparability of the entire number of hurricanes.

=== First improved graphic: Focus on the different world regions ===

[[Image:Verbessert3.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

=== Second improved graphic: Focus on the whole numbers of hurricanes ===

[[Image:Verbessert2.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

==Description of Improvements==

* Use of a 2-dimensional graph instead of a pseudo-3d graph
* Dividing the regions into seperate values.
* Use of different colors in the comparision between the entire number of hurricanes.
* Text in upper-and lowercase. We didn't use a serif font because on a computer screen a font without serifs is easier to read.

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

[Few, 2004] Stephen Few, Elegance Through Simplicity, intelligent enterprise, Oct. 16, 2004, http://www.intelligententerprise.com/showArticle.jhtml;jsessionid=N2ATDQWY5VYKSQSNDBGCKHSCJUMEKJVN?articleID=49400920

[Tufte, 1999] Edward R. Tufte: The Visual Display of Quantitative Information, Stanford University, Jan. 26, 1999, http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T21:50:27Z

UE-InfoVis0607 0026116: /* Principles of Graphical Integrity */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor Graphic==

===Principles of Graphical Integrity===
* '''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is ca.57% and the effect in the graphic about 55%, which results in a Lie Factor of nearly 1.The Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005], so my impression of the two graphs seems to be an optical illusion. Conclusion: Don't use graphform which tend to give the viewers wrong impressions of size differences.
* Because of the 3d-effect in the graphic the object in the background would be much more bigger in the foreground. Unfortunately the 3d-effect doesn't really come out at first sight and so the viewer compares the object directly.
* The numbers are not readable and the lines between the numbers are very irritating and confusing.
* As well is the colour gradient. The gradient goes in the smaler "3D Hurricane" from blue to black and in the bigger one reverse.
* a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
* A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
* The white lines forming squares (left) do not contain additional information.
* The background color has no additional information.
* It's not necessary to write "CAT. 4-5 Hurricanes" twice.
* In this case you could remove the numbers for the region, because they are unreadable.

'''Missing ink'''
* The ink doesn't change as the data changes.

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
* The "3D Hurricanes" take a lot of space. Its better to replace them with smaler representations of the data, e.g. bars.
* divide the regions in seperate

===Chartjunk===
* The white grid and the shadow in the background is used to imitate a 3d-effect. The 3d-effect is not needed in this graphic and so the grid should be junked.
* In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
* The text "bigger, faster, stronger, more" is not necessary. The graphic shows (or should show) that there are more Hurricanes than in the last years.

=== Design ===
* The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
* Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
* Although the data is about hurricans it doesn't suggets to use the shape of a hurrican in the graphic. It's better to use a horizontal graphic about 50% wider than tall.

* The type is not very clear or modest.
* The word "category" is not spelt out.

* The data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(151) shown in the same graphic.
* Some text in the graphic is written in uppercase only letters. That's difficult to read.

==Improved Graphic==

We made two improved graphics, the difference between them is that the first focuses on the comparability of the values in the different world regions and the second on the comparability of the entire number of hurricanes.

=== First improved graphic: Focus on the different world regions ===

[[Image:Verbessert3.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

=== Second improved graphic: Focus on the whole numbers of hurricanes ===

[[Image:Verbessert2.jpg|left|thumb|400px|none|Bigger, Faster, Stronger, More ... comparision of category 4-5 hurricans in different world regions 1975-1989 and 1990-2004]]

==Description of Improvements==

* Use of a 2-dimensional graph instead of a pseudo-3d graph
* Dividing the regions into seperate values.
* Use of different colors in the comparision between the entire number of hurricanes.
* Text in upper-and lowercase. We didn't use a serif font because on a computer screen a font without serifs is easier to read.

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

[Few, 2004] Stephen Few, Elegance Through Simplicity, intelligent enterprise, Oct. 16, 2004, http://www.intelligententerprise.com/showArticle.jhtml;jsessionid=N2ATDQWY5VYKSQSNDBGCKHSCJUMEKJVN?articleID=49400920

[Tufte, 1999] Edward R. Tufte: The Visual Display of Quantitative Information, Stanford University, Jan. 26, 1999, http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 1 - Data Density

2006-11-13T12:00:18Z

UE-InfoVis0607 0026116: /* Quotation */

== Quotation ==
{{Quotation|Data Density means the number of data points per area of the graph (that is the number of pixels in the display where number of pixels does not include the pixels in the window borders, menus, etc.).|[Brath, 1997]}}

{{Quotation|Number of Data Points means the number of discrete data values represented on screen at an instant.|[Barth, 1997]}}

== Charts and Graphs ==
The human eyes can detect large amount of different information within a small area, charts and graphs often ignore this ability.Data density gives you the opportunity to messure the information content of a graph or chart.
Therefor you need numerical values which relate the number of maximal at the same time presentable information elements to a display panel. [Edlinger, 2006]

Data density is also used to compute the cognitive complexity. [Edlinger, 2006] To do this you additonaly need the number of dimensions. [Dal Sasso Freitas, 2002]

In "The Visual Display of Quantitative Information" of Edward Tufte, a milestone in the theory of graph design, principles of graph design are discussed. One of them is:
Try to maximise the data density. [Hunt and Mashhoudy, 2002] and the size of the data matrix, within reason. As the volume of data increases, data measures must shrink and the graphics can be shrunk way down. [Tufte, 1999]

The maximum known density for a graph is 110,000 numbers/sq-inch for an astronomical graph. To compare, for most scientific journals we get about 50-200 numbers/sq-inch. [Smith, 2005]

Example for data densities include:
* 150 Mbits = human eye
* 8 Mbits = typical computer screen
* 25 Mbits = color slide
* 150 Mbits = large foldout map
* 28,000 Characters = Reference book
* 18,000 Characters = phone book
* 15,000 Characters = non-fiction
[Smith, 2005]

== Example ==
To better understand this term, we show two graphs and compute their data density. Consider the result of a survey in which the gender, height and weight were recorded for 92 students. The charts are 5.6cm by 7.4cm, an area of 41.4cm2. [Hunt and Mashhoudy, 2002]

[[Image:Data1.10.gif]]
[[Image:Data1.11.gif]]

Above you see a bar graph showing the breakdown into males and females.
According to the definition you've 2 (female and male) areas with 2 kind of information (gender and number of students), what makes 4 data points. By deviding this over the size of its area, you get the data density of this graph. [Barth, 1997]

number of data points = 4

data density = 4 / 41.4 = 0.1 (to 1 dp)
[Hunt and Mashhoudy, 2002]

The computation of the second example works similiar to the first one. The second graph shows a labeled XY-chart, which additonaly shows the relationship between height and weight of the students. As a result you've 3 kind of information on 92 areas, what results 276 data points. [Barth, 1997]

number of data points = 92 x 3 = 276

data density = 276 / 41.4 = 6.7 (to 1 dp)
[Hunt and Mashhoudy, 2002]

== References ==

[Barth, 1997] R. Barth. Metrics for effective information visualization. In ''Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis '97)'', pages 0–108, DC, USA, October 1997. IEEE Computer Society Washington

[Dal Sasso Freitas et al., 2002] Carla M. Dal Sasso Freitas, Paulo R. G. Luzzardi, Ricardo A. Cava, Marco A. A. Winckler, Marcelo S. Pimenta, Luciana P. Nedel. Evaluating Usability of Information Visualization Techniques. In ''Proceedings of 5th Symposium on Human Factors in Computer Systems (IHC)'', pages 10-11, Fortaleza, CE, 2002. Fortaleza:SBC

[Edlinger, 2006] Karl Edlinger, Informationsvisualisierung im Wissensmanagement – Eine Analyse unterschiedlicher Visualisierungstechniken auf ihre Eignung für das Wissensmanagement, Master's thesis, Fachhochschul-Studiengang Informationsberufe, Eisenstadt, 2006, 31-32.

[Hunt and Mashhoudy, 2002] Neville Hunt and Housh Mashhoudy, Discovering Important Statistical Concepts Using SpreadSheets. Created at: January 29, 2002. Retrieved at: October 28, 2006. http://home.ched.coventry.ac.uk/Volume/vol0/philosop.htm.

[Smith, 2005] Waynes Smith, Graphics and Web Design Based on Edward Tufte's Principles. Created at: January 17, 2005. Retrieved at: October 29, 2006. http://www.washington.edu/computing/training/560/zz-tufte.html

[Tufte, 1999] Edward R. Tufte, The Visual Display of Quantitative Information. Created at: January 26, 1999. Retrieved at: October 29, 2006. http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T11:55:22Z

UE-InfoVis0607 0026116: /* ..... */

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T10:44:15Z

UE-InfoVis0607 0026116: /* Principles of Graphical Integrity */

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T10:40:43Z

UE-InfoVis0607 0026116:

werde mich heute um eine verbesserte grafik kümmern, hab schon eine gute idee dafür.
[[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]]

noch was hab ich grad entdeckt: die grafik ist falsch! alle werte des zeitraumes 75-99 ergeben zusammen addiert 151 und nicht 171. kann das wer außer mir nochmal nachzählen und mich bestätigen oder widerlegen?
[[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]] 19:34, 12 November 2006 (CET)

ich komm auf 160.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

wie geht das? ich habe 1+10+16+23+36+65=151. [[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]]

hatte 18 statt 10, deins passt.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

danke.

eine frage: was ist uns wichtiger? dass die verschlechterungen in den einzelnen gebiete damals/heute vergleichen und der vergleich der gesamtsumme ein bissl weniger betont da steht, oder dass wir die gesamtsummen im vordergrund stehen haben, wodurch es schwerer wird, die entwicklungen in den einzelnen regionen zu vergleichen? [[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]]

ich bin dafür, dass der regionenvergleich wichtig ist und nicht der gesamtvergleich. wie kommt ihr auf 269, ich hab da 7 + 22 + 25 + 40 + 50 + 116 = 260??? [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]

es sind 49 im east pacific und nicht 40!

jo, das hab ich mir zurest auch gedacht,aber dann verworfen weil die zahlen einfach so extrem schlecht zu lesen sind.
Wegen der Grafik - meine erste Idee wäre einfachein Balkendiagramm: Pro Ort ein Balken, in dem dann noch eine Linie ist für das Vorjahr. [[User:UE-InfoVis0607 0325237|UE-InfoVis0607 0325237]]

also ich sehe zwei möglichkeiten das sinnvoll und simpel darzustellen:

variante 1: wie nici (0325237) geschrieben hat, 6 balken wobei jeder balken den vorgängerwert in entweder einem extrabalken dahinter (wie bei einem wahlergebnis, siehe [http://de.wikipedia.org/wiki/Bild:Nationalratswahl_2006_österreich.svg hier]), oder mittels linie drinnen hat.

variante 2: zwei balkendiagramme (für beide zeiträume) nebeneinander, wobei die balken jeweils eine zusammensetzung aus den 6 gebieten sind. zusätzlich so eine verbindungslinie damit man die steigerung der einzelnen gebiete sieht. ich bevorzuge dieses diagramm und werd mal einen entwurf posten.

[[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]] 22:47, 12 November 2006 (CET)

klingt auch gut, nur befürchte ich, dass das dann wieder zu kompliziert wird - dass man nicht gleich sieht was gemeint ist .... mal schaun [[User:UE-InfoVis0607 0325237|UE-InfoVis0607 0325237]]

so hab jetzt mal 3 bilder für euch. bin natürlich offen für jegliche verbesserungsvorschläge.

[[Image:Verbessert1a.jpg]]
[[Image:Verbessert1b.jpg]]
[[Image:Verbessert2.jpg]]

Mir gefallen alle drei gut, trotzdem ein paar punkte:

die unterschiedlichen farbschattierungen in der gleichen datenebene (northern indian, southwest, pacific, usw.) führen glaub ich zu unterschiedlich gewichteten daten, da es von der farbsättigung her aber ziemlich ähnliche farben sind, ist das kein allzu großes problem. trotzdem würd ich vorschlagen, nur eine farbe zu verwenden.

beim ersten vorschlag sehe ich den vergleich der jahre nicht heraus, und das denke ich ist doch sehr wichtig.

das zweite gefällt mir am besten, vielleicht aber auch nur weil ichs mir auch so vorgestellt habe :)

beim dritten kommt der gesamtvergleich wunderbar heraus, allerdings finde ich es für die repräsentation der daten besser, wenn die einzelnen gebiete direkt miteinander verglichen werden. allerdings wurde bei der orginalgrafik auch viel wert auf den gesamtvergleich gelegt.
ich kann halt mit diesem typ grafik nichts anfangen - vor allem auch nicht in zeitungen und im "echten leben"

aber schön sind sie alle drei :) [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]

noch eine anmerkung zu variante zwei: ich denke, die beschriftung der einzelnen regionen sollte direkt auf der x-achse passieren, und die legende sollte die zeiträume beschreiben. dh. mit zwei verschiedenen farben (hell/dunkel) für die zeiträume 75-89 und 90-04.
somit wäre die gesamtfarbenanzahl auf 2 minimiert. und diese beiden sollten meiner meinung nach nur helle und dunkle variation einer farbe sein, zb. dunkel/hellblau [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]

mir gefällt die dritte am besten, bei den ersten beiden fehlt mir die klar ersichtliche zeitraumaufteilung.
mit den im vorigen post beschriebenen änderungen wär das zweite auch gut. ist glaub ich eine geschmackssache ob einzel- od. gesamtvergleich höher gewichtet werden.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

Mir gefallen auch alle drei. hui. Ich wäre fast dafür zwei Verbesserungen der Grafik zu nehmen und das ganauso auch hinzuschreiben. Also das je nach dem was man damit zeigen will, die andere Darstellung besser geeignet ist.

Oder wir gehen jetzt her undn schauen uns die verschiedenen Faktoren an und argumentieren mit denen wieso eine der beiden ev. doch besser ist, aber ...

und eins noch: ja legenden sind nicht so gut
[[User:UE-InfoVis0607 0325237|UE-InfoVis0607 0325237]]

ja warum nicht, nehmen wir beide.
kurz zum text:"The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97."was genau mit diesem satz gemeint? den sollten wir anders formulieren.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

naja, lie factor berechent sich aus dem "effect in data" und "effect in graphic". bitte gerne anders formulieren.

bin auch für beide [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]

wenn ich den lie-factor ausrechne komm ich auf 68% für effect in data und ca. 60% effect in graphic. also ziemlich genau 1 beim lie-factor.is also wahrscheinlich nur eine optische täuschung das das rechte so viel grösser wirkt.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

hmm, ich bin mir nicht sicher ob meins stimmt. aber ich schreib mal trotzdem wie ich auf das ergebnis komme. habs in photoshop riengehaut und geschaut um wieviel ich das eine vergrößern muss damit ich zum anderen komme ->1,55. und der andere wert: 269-171/171 [[User:UE-InfoVis0607 0325237|UE-InfoVis0607 0325237]]

dann wäre effect in graphic 55% und effect in data 57%, weil 1,55 mal grösser bedeutet um 55% grösser.
einverstanden? ich werds mal ausbessern,ok.
[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T10:30:35Z

UE-InfoVis0607 0026116:

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T10:30:08Z

UE-InfoVis0607 0026116:

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:51:29Z

UE-InfoVis0607 0026116:

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:46:22Z

UE-InfoVis0607 0026116: /* Improved graphic */

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:46:11Z

UE-InfoVis0607 0026116: /* Arguments why it is a poor graphic */

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:45:54Z

UE-InfoVis0607 0026116: /* description of improvements */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor graphic==

===Principles of Graphical Integrity===
*'''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97.(BIN MIR NICHT SICHER) Its Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005]
*the numbers are not readable and the lines between the numbers are very irritating and confusing.
*a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
*A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
*The white lines forming squares (left) do not contain additional information.
*the background color has no additional information.
*It's not necessary to write "CAT. 4-5 Hurricanes" twice.
*in this case you could remove the numbers for the region, because they are unreadable.

'''Missing ink'''
*The ink doesn't change as the data changes.

===Visual Clutter===

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
*

===Chartjunk===

=== ..... ===
*The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
*Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
*In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
*The word "category" is not spelt out.

*multivariate
*efficent in interpretation?
*often narrative in content, showing shifts in the relationship between variables as the index variable change
*the data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(160) shown in the same graphic.

==Improved graphic==

==Description of Improvements==

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:45:39Z

UE-InfoVis0607 0026116: /* improved graphic */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor graphic==

===Principles of Graphical Integrity===
*'''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97.(BIN MIR NICHT SICHER) Its Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005]
*the numbers are not readable and the lines between the numbers are very irritating and confusing.
*a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
*A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
*The white lines forming squares (left) do not contain additional information.
*the background color has no additional information.
*It's not necessary to write "CAT. 4-5 Hurricanes" twice.
*in this case you could remove the numbers for the region, because they are unreadable.

'''Missing ink'''
*The ink doesn't change as the data changes.

===Visual Clutter===

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
*

===Chartjunk===

=== ..... ===
*The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
*Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
*In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
*The word "category" is not spelt out.

*multivariate
*efficent in interpretation?
*often narrative in content, showing shifts in the relationship between variables as the index variable change
*the data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(160) shown in the same graphic.

==Improved graphic==

==description of improvements==

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:44:55Z

UE-InfoVis0607 0026116: /* Principles of Graphical Integrity */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor graphic==

===Principles of Graphical Integrity===
*'''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97.(BIN MIR NICHT SICHER) Its Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005]
*the numbers are not readable and the lines between the numbers are very irritating and confusing.
*a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
*A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
*The white lines forming squares (left) do not contain additional information.
*the background color has no additional information.
*It's not necessary to write "CAT. 4-5 Hurricanes" twice.
*in this case you could remove the numbers for the region, because they are unreadable.

'''Missing ink'''
*The ink doesn't change as the data changes.

===Visual Clutter===

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
*

===Chartjunk===

=== ..... ===
*The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
*Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
*In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
*The word "category" is not spelt out.

*multivariate
*efficent in interpretation?
*often narrative in content, showing shifts in the relationship between variables as the index variable change
*the data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(160) shown in the same graphic.

==improved graphic==
==description of improvements==

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:43:49Z

UE-InfoVis0607 0026116: /* Data-Ink Ratio */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor graphic==

===Principles of Graphical Integrity===
*'''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97.(BIN MIR NICHT SICHER) Its Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005]
*the numbers are not readable and the lines between the numbers are not needed and very irritating.
*a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
*A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
*The white lines forming squares (left) do not contain additional information.
*the background color has no additional information.
*It's not necessary to write "CAT. 4-5 Hurricanes" twice.
*in this case you could remove the numbers for the region, because they are unreadable.

'''Missing ink'''
*The ink doesn't change as the data changes.

===Visual Clutter===

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
*

===Chartjunk===

=== ..... ===
*The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
*Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
*In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
*The word "category" is not spelt out.

*multivariate
*efficent in interpretation?
*often narrative in content, showing shifts in the relationship between variables as the index variable change
*the data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(160) shown in the same graphic.

==improved graphic==
==description of improvements==

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:43:30Z

UE-InfoVis0607 0026116: /* Data-Ink Ratio */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor graphic==

===Principles of Graphical Integrity===
*'''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97.(BIN MIR NICHT SICHER) Its Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005]
*the numbers are not readable and the lines between the numbers are not needed and very irritating.
*a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
*A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
*The white lines forming squares (left) do not contain additional information.
*the background color has no additional information.
*It's not necessary to write "CAT. 4-5 Hurricanes" twice.
*in this case you could remove the numbers for the region, because they are unreadable.

'''Missing ink'''
*The ink don't change as the data change.

===Visual Clutter===

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
*

===Chartjunk===

=== ..... ===
*The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
*Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
*In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
*The word "category" is not spelt out.

*multivariate
*efficent in interpretation?
*often narrative in content, showing shifts in the relationship between variables as the index variable change
*the data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(160) shown in the same graphic.

==improved graphic==
==description of improvements==

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:37:20Z

UE-InfoVis0607 0026116:

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-13T09:36:56Z

UE-InfoVis0607 0026116:

werde mich heute um eine verbesserte grafik kümmern, hab schon eine gute idee dafür.
[[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]]

noch was hab ich grad entdeckt: die grafik ist falsch! alle werte des zeitraumes 75-99 ergeben zusammen addiert 151 und nicht 171. kann das wer außer mir nochmal nachzählen und mich bestätigen oder widerlegen?
[[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]] 19:34, 12 November 2006 (CET)

ich komm auf 160.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

wie geht das? ich habe 1+10+16+23+36+65=151. [[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]]

hatte 18 statt 10, deins passt.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

danke.

eine frage: was ist uns wichtiger? dass die verschlechterungen in den einzelnen gebiete damals/heute vergleichen und der vergleich der gesamtsumme ein bissl weniger betont da steht, oder dass wir die gesamtsummen im vordergrund stehen haben, wodurch es schwerer wird, die entwicklungen in den einzelnen regionen zu vergleichen? [[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]]

ich bin dafür, dass der regionenvergleich wichtig ist und nicht der gesamtvergleich. wie kommt ihr auf 269, ich hab da 7 + 22 + 25 + 40 + 50 + 116 = 260??? [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]

es sind 49 im east pacific und nicht 40!

jo, das hab ich mir zurest auch gedacht,aber dann verworfen weil die zahlen einfach so extrem schlecht zu lesen sind.
Wegen der Grafik - meine erste Idee wäre einfachein Balkendiagramm: Pro Ort ein Balken, in dem dann noch eine Linie ist für das Vorjahr. [[User:UE-InfoVis0607 0325237|UE-InfoVis0607 0325237]]

also ich sehe zwei möglichkeiten das sinnvoll und simpel darzustellen:

variante 1: wie nici (0325237) geschrieben hat, 6 balken wobei jeder balken den vorgängerwert in entweder einem extrabalken dahinter (wie bei einem wahlergebnis, siehe [http://de.wikipedia.org/wiki/Bild:Nationalratswahl_2006_österreich.svg hier]), oder mittels linie drinnen hat.

variante 2: zwei balkendiagramme (für beide zeiträume) nebeneinander, wobei die balken jeweils eine zusammensetzung aus den 6 gebieten sind. zusätzlich so eine verbindungslinie damit man die steigerung der einzelnen gebiete sieht. ich bevorzuge dieses diagramm und werd mal einen entwurf posten.

[[User:UE-InfoVis0607 0325529|UE-InfoVis0607 0325529]] 22:47, 12 November 2006 (CET)

klingt auch gut, nur befürchte ich, dass das dann wieder zu kompliziert wird - dass man nicht gleich sieht was gemeint ist .... mal schaun [[User:UE-InfoVis0607 0325237|UE-InfoVis0607 0325237]]

so hab jetzt mal 3 bilder für euch. bin natürlich offen für jegliche verbesserungsvorschläge.

[[Image:Verbessert1a.jpg]]
[[Image:Verbessert1b.jpg]]
[[Image:Verbessert2.jpg]]

Mir gefallen alle drei gut, trotzdem ein paar punkte:

die unterschiedlichen farbschattierungen in der gleichen datenebene (northern indian, southwest, pacific, usw.) führen glaub ich zu unterschiedlich gewichteten daten, da es von der farbsättigung her aber ziemlich ähnliche farben sind, ist das kein allzu großes problem. trotzdem würd ich vorschlagen, nur eine farbe zu verwenden.

beim ersten vorschlag sehe ich den vergleich der jahre nicht heraus, und das denke ich ist doch sehr wichtig.

das zweite gefällt mir am besten, vielleicht aber auch nur weil ichs mir auch so vorgestellt habe :)

beim dritten kommt der gesamtvergleich wunderbar heraus, allerdings finde ich es für die repräsentation der daten besser, wenn die einzelnen gebiete direkt miteinander verglichen werden. allerdings wurde bei der orginalgrafik auch viel wert auf den gesamtvergleich gelegt.
ich kann halt mit diesem typ grafik nichts anfangen - vor allem auch nicht in zeitungen und im "echten leben"

aber schön sind sie alle drei :) [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]

noch eine anmerkung zu variante zwei: ich denke, die beschriftung der einzelnen regionen sollte direkt auf der x-achse passieren, und die legende sollte die zeiträume beschreiben. dh. mit zwei verschiedenen farben (hell/dunkel) für die zeiträume 75-89 und 90-04.
somit wäre die gesamtfarbenanzahl auf 2 minimiert. und diese beiden sollten meiner meinung nach nur helle und dunkle variation einer farbe sein, zb. dunkel/hellblau [[User:UE-InfoVis0607 0226388|UE-InfoVis0607 0226388]]
mir gefällt die dritte am besten, bei den ersten beiden fehlt mir die klar ersichtliche zeitraumaufteilung.
mit den im vorigen post beschriebenen änderungen wär das zweite auch gut. ist glaub ich eine geschmackssache ob einzel- od. gesamtvergleich höher gewichtet werden.[[User:UE-InfoVis0607 0026116|UE-InfoVis0607 0026116]]

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-12T21:29:37Z

UE-InfoVis0607 0026116: /* ..... */

== Poor Graphic ==

[[Image:Seed06storms.jpg|none|thumb|600px|none|State of the Planet - Bigger, Faster, Stronger, More]]

==Arguments why it is a poor graphic==

===Principles of Graphical Integrity===
*'''Lie Factor''': The right graphic seems do be much larger then the left. It should be only 1,5 times larger then the other one based on the data. The size of the effect in data is 5,7% and the effect in the graphic about 1,55%, which results a high Lie Factor of 2,97.(BIN MIR NICHT SICHER) Its Lie Factor should have a value between 0.95 and 1.05. [Friendly, 2005]
*the numbers are not readable and the lines between the numbers are not needed and very irritating.
*a pseudo 3-dimensional graphic is used to show 1-dimensional data

===Data-Ink Ratio===
This graph has a low Data-Ink Ratio. To maximise the data-ink ratio we have to reduce the non-data ink and enhance the data ink.

'''How much of the ink could be removed without loss of information?'''
*A large share of ink is used to show the pseudo 3-dimensional graphic and their shadows.
*The white lines forming squares (left) do not contain additional information.
*It's not necessary to wirte two times "CAT. 4-5 Hurricanes".
*

'''Missing ink'''
*The ink don't change as the data change.
*

===Visual Clutter===

===Data Density===
The Data Density of this graphic is very low. We should maximise it and the size of the data matrix, within reason.
* The graphic is very big and should be shrunk way down.
*

===Chartjunk===

=== ..... ===
*The graphic isn't very comparative. The smaler "3D Hurricane" appears to be just farther behind in time and position.
*Graphical elegance is often found in simplicity of design and complexity of data, but in this graphic the graphical design is more complex than the visualised data.
*In the text a reason for increased hurricans is mentioned, but in the shown data is no evidence for that, because there is no related data.
*The word "category" is not spelt out.

*multivariate
*efficent in interpretation?
*often narrative in content, showing shifts in the relationship between variables as the index variable change
*the data is wrong, the sum of the hurricans(171) as shown in the left graphic does not equal the summed up hurricans from the regions(160) shown in the same graphic.

==improved graphic==
==description of improvements==

== References ==
[Friendly, 2005] Michael Friendly, The Lie Factor, Department of Mathematics and Statistics, Toronto, Last updated: February 23, 2002, Retrieved at: Nov 17, 2006, http://www.math.yorku.ca/SCS/Gallery/lie-factor.html

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-12T18:57:11Z

UE-InfoVis0607 0026116:

Teaching talk:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 2

2006-11-12T18:47:27Z

UE-InfoVis0607 0026116:

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 1 - Data Density

2006-11-12T17:50:52Z

UE-InfoVis0607 0026116: /* Charts and Graphs */

== Quotation ==
{{Quotation|Data Density means the number of data points per area of the graph (that is the number of pixels in the display where number of pixels does not include the pixels in the window borders, menus, etc.).}} [Brath, 1997]

{{Quotation|Number of Data Points means the number of discrete data values represented on screen at an instant.}} [Barth, 1997]

== Charts and Graphs ==
The human eyes can detect large amount of different information within a small area, charts and graphs often ignore this ability.Data density gives you the opportunity to messure the information content of a graph or chart.
Therefor you need numerical values which relate the number of maximal at the same time presentable information elements to a display panel. [Edlinger, 2006]

Data density is also used to compute the cognitive complexity. [Edlinger, 2006] To do this you additonaly need the number of dimensions. [Dal Sasso Freitas, 2002]

In "The Visual Display of Quantitative Information" of Edward Tufte, a milestone in the theory of graph design, principles of graph design are discussed. One of them is:
Try to maximise the data density. [Hunt and Mashhoudy, 2002] and the size of the data matrix, within reason. As the volume of data increases, data measures must shrink and the graphics can be shrunk way down. [Tufte, 1999]

The maximum known density for a graph is 110,000 numbers/sq-inch for an astronomical graph. To compare, for most scientific journals we get about 50-200 numbers/sq-inch. [Smith, 2005]

Example for data densities include:
* 150 Mbits = human eye
* 8 Mbits = typical computer screen
* 25 Mbits = color slide
* 150 Mbits = large foldout map
* 28,000 Characters = Reference book
* 18,000 Characters = phone book
* 15,000 Characters = non-fiction
[Smith, 2005]

== Example ==
To better understand this term, we show two graphs and compute their data density. Consider the result of a survey in which the gender, height and weight were recorded for 92 students. The charts are 5.6cm by 7.4cm, an area of 41.4cm2. [Hunt and Mashhoudy, 2002]

[[Image:Data1.10.gif]]
[[Image:Data1.11.gif]]

Above you see a bar graph showing the breakdown into males and females.
According to the definition you've 2 (female and male) areas with 2 kind of information (gender and number of students), what makes 4 data points. By deviding this over the size of its area, you get the data density of this graph. [Barth, 1997]

number of data points = 4

data density = 4 / 41.4 = 0.1 (to 1 dp)
[Hunt and Mashhoudy, 2002]

The computation of the second example works similiar to the first one. The second graph shows a labeled XY-chart, which additonaly shows the relationship between height and weight of the students. As a result you've 3 kind of information on 92 areas, what results 276 data points. [Barth, 1997]

number of data points = 92 x 3 = 276

data density = 276 / 41.4 = 6.7 (to 1 dp)
[Hunt and Mashhoudy, 2002]

== References ==

[Barth, 1997] R. Barth. Metrics for effective information visualization. In ''Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis '97)'', pages 0–108, DC, USA, October 1997. IEEE Computer Society Washington

[Dal Sasso Freitas et al., 2002] Carla M. Dal Sasso Freitas, Paulo R. G. Luzzardi, Ricardo A. Cava, Marco A. A. Winckler, Marcelo S. Pimenta, Luciana P. Nedel. Evaluating Usability of Information Visualization Techniques. In ''Proceedings of 5th Symposium on Human Factors in Computer Systems (IHC)'', pages 10-11, Fortaleza, CE, 2002. Fortaleza:SBC

[Edlinger, 2006] Karl Edlinger, Informationsvisualisierung im Wissensmanagement – Eine Analyse unterschiedlicher Visualisierungstechniken auf ihre Eignung für das Wissensmanagement, Master's thesis, Fachhochschul-Studiengang Informationsberufe, Eisenstadt, 2006, 31-32.

[Hunt and Mashhoudy, 2002] Neville Hunt and Housh Mashhoudy, Discovering Important Statistical Concepts Using SpreadSheets. Created at: January 29, 2002. Retrieved at: October 28, 2006. http://home.ched.coventry.ac.uk/Volume/vol0/philosop.htm.

[Smith, 2005] Waynes Smith, Graphics and Web Design Based on Edward Tufte's Principles. Created at: January 17, 2005. Retrieved at: October 29, 2006. http://www.washington.edu/computing/training/560/zz-tufte.html

[Tufte, 1999] Edward R. Tufte, The Visual Display of Quantitative Information. Created at: January 26, 1999. Retrieved at: October 29, 2006. http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 1 - Data Density

2006-11-12T17:37:57Z

UE-InfoVis0607 0026116: /* Example */

== Quotation ==
{{Quotation|Data Density means the number of data points per area of the graph (that is the number of pixels in the display where number of pixels does not include the pixels in the window borders, menus, etc.).}} [Brath, 1997]

{{Quotation|Number of Data Points means the number of discrete data values represented on screen at an instant.}} [Barth, 1997]

== Charts and Graphs ==
While the human eye has the ability to detect large amounts of information within a small area, statistical graphs often fail to take advantage of this ability. To messure the information content of a graph he definses the data density. [Hunt and Mashhoudy, 2002] Therefor you need numerical values which relate the number of maximal at the same time presentable information elements to a display panel. [Edlinger, 2006]

Data density is also used to compute the cognitive complexity. [Edlinger, 2006] To do this you additonaly need the number of dimensions. [Dal Sasso Freitas, 2002]

In "The Visual Display of Quantitative Information" of Edward Tufte, a milestone in the theory of graph design, principles of graph design are discussed. One of them is:
Try to maximise the data density. [Hunt and Mashhoudy, 2002] and the size of the data matrix, within reason. As the volume of data increases, data measures must shrink and the graphics can be shrunk way down. [Tufte, 1999]

Usually there is a minimum amount, which should always be less than the data density. If the data density falls below this value 0 is used as a result, otherwise 1. [Edlinger, 2006]

The maximum known density for a graph is 110,000 numbers/sq-inch for an astronomical graph. To compare, for most scientific journals we get about 50-200 numbers/sq-inch. [Smith, 2005]

Example for data densities include:
* 150 Mbits = human eye
* 8 Mbits = typical computer screen
* 25 Mbits = color slide
* 150 Mbits = large foldout map
* 28,000 Characters = Reference book
* 18,000 Characters = phone book
* 15,000 Characters = non-fiction
[Smith, 2005]

== Example ==
To better understand this term, we show two graphs and compute their data density. Consider the result of a survey in which the gender, height and weight were recorded for 92 students. The charts are 5.6cm by 7.4cm, an area of 41.4cm2. [Hunt and Mashhoudy, 2002]

[[Image:Data1.10.gif]]
[[Image:Data1.11.gif]]

Above you see a bar graph showing the breakdown into males and females.
According to the definition you've 2 (female and male) areas with 2 kind of information (gender and number of students), what makes 4 data points. By deviding this over the size of its area, you get the data density of this graph. [Barth, 1997]

number of data points = 4

data density = 4 / 41.4 = 0.1 (to 1 dp)
[Hunt and Mashhoudy, 2002]

The computation of the second example works similiar to the first one. The second graph shows a labeled XY-chart, which additonaly shows the relationship between height and weight of the students. As a result you've 3 kind of information on 92 areas, what results 276 data points. [Barth, 1997]

number of data points = 92 x 3 = 276

data density = 276 / 41.4 = 6.7 (to 1 dp)
[Hunt and Mashhoudy, 2002]

== References ==

[Barth, 1997] R. Barth. Metrics for effective information visualization. In ''Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis '97)'', pages 0–108, DC, USA, October 1997. IEEE Computer Society Washington

[Dal Sasso Freitas et al., 2002] Carla M. Dal Sasso Freitas, Paulo R. G. Luzzardi, Ricardo A. Cava, Marco A. A. Winckler, Marcelo S. Pimenta, Luciana P. Nedel. Evaluating Usability of Information Visualization Techniques. In ''Proceedings of 5th Symposium on Human Factors in Computer Systems (IHC)'', pages 10-11, Fortaleza, CE, 2002. Fortaleza:SBC

[Edlinger, 2006] Karl Edlinger, Informationsvisualisierung im Wissensmanagement – Eine Analyse unterschiedlicher Visualisierungstechniken auf ihre Eignung für das Wissensmanagement, Master's thesis, Fachhochschul-Studiengang Informationsberufe, Eisenstadt, 2006, 31-32.

[Hunt and Mashhoudy, 2002] Neville Hunt and Housh Mashhoudy, Discovering Important Statistical Concepts Using SpreadSheets. Created at: January 29, 2002. Retrieved at: October 28, 2006. http://home.ched.coventry.ac.uk/Volume/vol0/philosop.htm.

[Smith, 2005] Waynes Smith, Graphics and Web Design Based on Edward Tufte's Principles. Created at: January 17, 2005. Retrieved at: October 29, 2006. http://www.washington.edu/computing/training/560/zz-tufte.html

[Tufte, 1999] Edward R. Tufte, The Visual Display of Quantitative Information. Created at: January 26, 1999. Retrieved at: October 29, 2006. http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

File:Chart.png

2006-11-12T17:32:43Z

UE-InfoVis0607 0026116: /* Quelle */

== Beschreibung ==
InfoZoom Graphics Report
== Copyright-Status ==

== Quelle ==
[Xie, 2006] Zaixian Xie, Review on InfoZoom. Worcester Polytechnic Insitute: Course Data Visualization. Created at: 1. Feb 2006. Retrieved at: 30. Oct 2006. http://users.wpi.edu/~xiezx/courses/cs525d/infozoom/infozoom.htm

File:Overview mode.gif

2006-11-12T17:32:15Z

UE-InfoVis0607 0026116: /* Quelle */

== Beschreibung ==
InfoZoom Overview mode
== Copyright-Status ==

== Quelle ==
[Spenke, 2000] Michael Spenke and Christian Beilken. InfoZoom - Analysing Formula One Racing Results With an Interactive Data Mining and Visualisation Tool. In ''Second Conference on Data Mining 2000'', pages 1-11, Cambridge, UK, July 2000. http://www.ifs.tuwien.ac.at/~silvia/wien/vu-infovis/articles/spenke00infozoom.pdf

File:Compressed table.gif

2006-11-12T17:31:57Z

UE-InfoVis0607 0026116: /* Quelle */

== Beschreibung ==
InfoZoom Compressed Table mode
== Copyright-Status ==

== Quelle ==
[Spenke, 2000] Michael Spenke and Christian Beilken. InfoZoom - Analysing Formula One Racing Results With an Interactive Data Mining and Visualisation Tool. In ''Second Conference on Data Mining 2000'', pages 1-11, Cambridge, UK, July 2000. http://www.ifs.tuwien.ac.at/~silvia/wien/vu-infovis/articles/spenke00infozoom.pdf

File:Wide table.gif

2006-11-12T17:31:37Z

UE-InfoVis0607 0026116: /* Quelle */

== Beschreibung ==
InfoZoom Wide Table mode
== Copyright-Status ==

== Quelle ==
[Spenke, 2000] Michael Spenke and Christian Beilken. InfoZoom - Analysing Formula One Racing Results With an Interactive Data Mining and Visualisation Tool. In ''Second Conference on Data Mining 2000'', pages 1-11, Cambridge, UK, July 2000. http://www.ifs.tuwien.ac.at/~silvia/wien/vu-infovis/articles/spenke00infozoom.pdf

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 1 - InfoZoom

2006-11-12T17:29:23Z

UE-InfoVis0607 0026116: /* Quotation */

== Quotation ==

{{Definition|InfoZoom is an application, developed by the company humanIT a spin-off of Fraunhofer Society, for analyzing and displaying large amounts of data in an easy-to-understand format.}}

=== History ===

InfoZoom evolved from another system called FOCUS, a focus+context table layout [Kort, 2004].

== Description ==

=== Overview ===

Primarily, InfoZoom was developed for manipulating tables. By default it uses parallel bargrams for larger sizes of data sets and special focus+context tables for smaller ones [Wittenburg et al, 2001]. So it is possible to present information in table or distribution oriented formats.

When displaying relational data tables, InfoZoom uses one row for each record. This data can be sorted sequential, hierarchical or in any other way the user wants it to be sorted. If necessary, the user can "zoom" into the information. This happens for example by doubleclicking on attribute values or sets/ranges of values [Mark et al, 2003]. This "zooming" can also be done by selecting columns of a row interactively and then promoting them into qualified records. All other records except those marked as qualified records are hidden instantly, enabling a "zoomed" view on the information.

{{Quotation|The primary interactions supported in InfoZoom are simple selection and filtering, accomplished with a single mouse click on a value bin. Interesting graphical renderings of values are included. Animation is used to transition between set sizes.| [Wittenburg et al, 2001]}}

=== Data source ===

The user can import data either with the import tool for Excel or text files or use an ODBC or OLE DB connection.

=== View Modes ===

InfoZoom offers three modes for data analysis: wide table, compressed table and overview

==== Wide Table ====

The wide table looks like a table we know from spreadsheets or database tables. Every attribute is shown as a row and the objects as columns. If the windows is not big enough to show all records or attributes you can scroll vertically or horizontally

[[Image:Wide_table.gif|200px]]

==== Compressed Table ====

The compressed table shows all objects pushed together to fit on the screen. Numerical values are represented by the vertical position of dots. The data can be sorted by individual attributes in ascending or descendig order.

[[Image:Compressed_table.gif|200px]]

==== Overview Mode ====

The overview mode allows that the content of each attributes row are sorted independently. The identical values are placed together in a cell. The width of the cell represents the frequency of this value.

[[Image:Overview_mode.gif|200px]]

=== Filter ===

With filter the user can search data or find patterns in a dataset. You can search after a specific value or value range on an attribute.

=== Derived Attribute ===

The user can combine attributes with derived attributes. The new value is calculated from the values of one or two already existing attributes.

=== Graphics Report ===

Graphic Reports can be created with different styles, e.g. Pie, Donut, Bar, Horiz Bar, Line, Area, Point, Bubble, Volume, ...

[[Image:Chart.png|200px]]

== Use ==

InfoZoom is often used for Visual Data Mining (VDM), a process which aims at the recognition of patterns in data by an human expert. VDM makes use of an technique called focus+context. Here, it's possible for the expert to show focused and contextual information at the same time. [Kort, 2004]

== References ==

[humanIT, 2006] Author unknown, InfoZoom. humanIT Software GmbH. Retrieved at: 30 Oct 2006. http://www.infozoom.com/enu/infozoom/index.htm

[Kort, 2004] Alexander Kort. Visual Data Mining and Zoomable Interfaces. In ''Proceedings of the 9th international conference on Intelligent user interface''. pages 274-276, New York, NY, 2004. ACM Press

[Mark et al, 2003] Gloria Mark, Keri Carpenter, Alfred Kobsa. Are There Benefits in Seeing Double? A Study of Collaborative Information Visualization. In ''CHI '03 extended abstracts on Human factors in computing systems'', pages 840-841, Irvine, CA, April 2003. ACM Press

[Spenke, 2000] Michael Spenke and Christian Beilken. InfoZoom - Analysing Formula One Racing Results With an Interactive Data Mining and Visualisation Tool. In ''Second Conference on Data Mining 2000'', pages 1-11, Cambridge, UK, July 2000. http://www.ifs.tuwien.ac.at/~silvia/wien/vu-infovis/articles/spenke00infozoom.pdf

[Wittenburg et al, 2001] Kent Wittenburg, Tom Lanning, Michael Heinrichs, Michael Stanton. Parallel bargrams for consumer-based information exploration and choice. In ''Proceedings of the 14th annual ACM symposium on User interface software and technology'', pages 51-60, Waltham, MA, November 2001. ACM Press

[Xie, 2006] Zaixian Xie, Review on InfoZoom. Worcester Polytechnic Insitute: Course Data Visualization. Created at: 1. Feb 2006. Retrieved at: 30. Oct 2006. http://users.wpi.edu/~xiezx/courses/cs525d/infozoom/infozoom.htm

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 1 - Data Density

2006-11-12T17:05:05Z

UE-InfoVis0607 0026116: /* Charts and Graphs */

== Quotation ==
{{Quotation|Data Density means the number of data points per area of the graph (that is the number of pixels in the display where number of pixels does not include the pixels in the window borders, menus, etc.).}} [Brath, 1997]

{{Quotation|Number of Data Points means the number of discrete data values represented on screen at an instant.}} [Barth, 1997]

== Charts and Graphs ==
While the human eye has the ability to detect large amounts of information within a small area, statistical graphs often fail to take advantage of this ability. To messure the information content of a graph he definses the data density. [Hunt and Mashhoudy, 2002] Therefor you need numerical values which relate the number of maximal at the same time presentable information elements to a display panel. [Edlinger, 2006]

Data density is also used to compute the cognitive complexity. [Edlinger, 2006] To do this you additonaly need the number of dimensions. [Dal Sasso Freitas, 2002]

In "The Visual Display of Quantitative Information" of Edward Tufte, a milestone in the theory of graph design, principles of graph design are discussed. One of them is:
Try to maximise the data density. [Hunt and Mashhoudy, 2002] and the size of the data matrix, within reason. As the volume of data increases, data measures must shrink and the graphics can be shrunk way down. [Tufte, 1999]

Usually there is a minimum amount, which should always be less than the data density. If the data density falls below this value 0 is used as a result, otherwise 1. [Edlinger, 2006]

The maximum known density for a graph is 110,000 numbers/sq-inch for an astronomical graph. To compare, for most scientific journals we get about 50-200 numbers/sq-inch. [Smith, 2005]

Example for data densities include:
* 150 Mbits = human eye
* 8 Mbits = typical computer screen
* 25 Mbits = color slide
* 150 Mbits = large foldout map
* 28,000 Characters = Reference book
* 18,000 Characters = phone book
* 15,000 Characters = non-fiction
[Smith, 2005]

== Example ==
To better undrestand this term, we show two graphs and compute their data density. Consider the result of a survey in which the gender, height and weight were recordet for 92 students. The charts are 5.6cm by 7.4cm, an area of 41.4cm2. [Hunt and Mashhoudy, 2002]

[[Image:Data1.10.gif]]
[[Image:Data1.11.gif]]

Above you see a bar graph showing the breakdown into males and females.
According the definition you've 2 (female and male) areas with 2 kind of information (gender and number of students), what makes 4 data points. By deviding this over the size of its area, you get the data tensity of this graph. [Barth, 1997]

number of data points = 4

data density = 4 / 41.4 = 0.1 (to 1 dp)
[Hunt and Mashhoudy, 2002]

The computation of the second example works similiar to the first one. The second graph shows a labelled XY-chart, which additonaly shows the relationship between height and weight of the students. As a result you've 3 kind of information on 92 areas, what results 276 data points. [Barth, 1997]

number of data points = 92 x 3 = 276

data density = 276 / 41.4 = 6.7 (to 1 dp)
[Hunt and Mashhoudy, 2002]

== References ==

[Barth, 1997] R. Barth. Metrics for effective information visualization. In ''Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis '97)'', pages 0–108, DC, USA, October 1997. IEEE Computer Society Washington

[Dal Sasso Freitas et al., 2002] Carla M. Dal Sasso Freitas, Paulo R. G. Luzzardi, Ricardo A. Cava, Marco A. A. Winckler, Marcelo S. Pimenta, Luciana P. Nedel. Evaluating Usability of Information Visualization Techniques. In ''Proceedings of 5th Symposium on Human Factors in Computer Systems (IHC)'', pages 10-11, Fortaleza, CE, 2002. Fortaleza:SBC

[Edlinger, 2006] Karl Edlinger, Informationsvisualisierung im Wissensmanagement – Eine Analyse unterschiedlicher Visualisierungstechniken auf ihre Eignung für das Wissensmanagement, Master's thesis, Fachhochschul-Studiengang Informationsberufe, Eisenstadt, 2006, 31-32.

[Hunt and Mashhoudy, 2002] Neville Hunt and Housh Mashhoudy, Discovering Important Statistical Concepts Using SpreadSheets. Created at: January 29, 2002. Retrieved at: October 28, 2006. http://home.ched.coventry.ac.uk/Volume/vol0/philosop.htm.

[Smith, 2005] Waynes Smith, Graphics and Web Design Based on Edward Tufte's Principles. Created at: January 17, 2005. Retrieved at: October 29, 2006. http://www.washington.edu/computing/training/560/zz-tufte.html

[Tufte, 1999] Edward R. Tufte, The Visual Display of Quantitative Information. Created at: January 26, 1999. Retrieved at: October 29, 2006. http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html

Teaching:TUW - UE InfoVis WS 2006/07 - Gruppe 03 - Aufgabe 1 - Data Density

2006-11-12T17:03:49Z

UE-InfoVis0607 0026116: /* Quotation */

== Quotation ==
{{Quotation|Data Density means the number of data points per area of the graph (that is the number of pixels in the display where number of pixels does not include the pixels in the window borders, menus, etc.).}} [Brath, 1997]

{{Quotation|Number of Data Points means the number of discrete data values represented on screen at an instant.}} [Barth, 1997]

== Charts and Graphs ==
While the human eye has the ability to detect large amounts of information within a small area, statistical graphs often fail to take advantage of this ability. To messure the information content of a graph he definses the data density. [Hunt and Mashhoudy, 2002] Therefor you need numerical values which relate the number of maximal at the same time presentable information elements to a display panel. [Edlinger, 2006]

Data density is also used to compute the cognitive complexity. [Edlinger, 2006] To do this you additonaly need the number of dimensions. [Dal Sasso Freitas, 2002]

In "The Visual Display of Graphical Information" of Edward Tufte, a milestone in the theory of graph design, principles of graph design are discussed. One of them is:
Try to maximise the data density. [Hunt and Mashhoudy, 2002] and the size of the data matrix, within reason. As the volume of data increases, data measures must shrink and the graphics can be shrunk way down. [Tufte, 1999]

Usually there is a minimum amount, which should always be less than the data density. If the data density falls below this value 0 is used as a result, otherwise 1. [Edlinger, 2006]

The maximum known density for a graph is 110,000 numbers/sq-inch for an astronomical graph. To compare, for most scientific journals we get about 50-200 numbers/sq-inch. [Smith, 2005]

Example for data densities include:
* 150 Mbits = human eye
* 8 Mbits = typical computer screen
* 25 Mbits = color slide
* 150 Mbits = large foldout map
* 28,000 Characters = Reference book
* 18,000 Characters = phone book
* 15,000 Characters = non-fiction
[Smith, 2005]

== Example ==
To better undrestand this term, we show two graphs and compute their data density. Consider the result of a survey in which the gender, height and weight were recordet for 92 students. The charts are 5.6cm by 7.4cm, an area of 41.4cm2. [Hunt and Mashhoudy, 2002]

[[Image:Data1.10.gif]]
[[Image:Data1.11.gif]]

Above you see a bar graph showing the breakdown into males and females.
According the definition you've 2 (female and male) areas with 2 kind of information (gender and number of students), what makes 4 data points. By deviding this over the size of its area, you get the data tensity of this graph. [Barth, 1997]

number of data points = 4

data density = 4 / 41.4 = 0.1 (to 1 dp)
[Hunt and Mashhoudy, 2002]

The computation of the second example works similiar to the first one. The second graph shows a labelled XY-chart, which additonaly shows the relationship between height and weight of the students. As a result you've 3 kind of information on 92 areas, what results 276 data points. [Barth, 1997]

number of data points = 92 x 3 = 276

data density = 276 / 41.4 = 6.7 (to 1 dp)
[Hunt and Mashhoudy, 2002]

== References ==

[Barth, 1997] R. Barth. Metrics for effective information visualization. In ''Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis '97)'', pages 0–108, DC, USA, October 1997. IEEE Computer Society Washington

[Dal Sasso Freitas et al., 2002] Carla M. Dal Sasso Freitas, Paulo R. G. Luzzardi, Ricardo A. Cava, Marco A. A. Winckler, Marcelo S. Pimenta, Luciana P. Nedel. Evaluating Usability of Information Visualization Techniques. In ''Proceedings of 5th Symposium on Human Factors in Computer Systems (IHC)'', pages 10-11, Fortaleza, CE, 2002. Fortaleza:SBC

[Edlinger, 2006] Karl Edlinger, Informationsvisualisierung im Wissensmanagement – Eine Analyse unterschiedlicher Visualisierungstechniken auf ihre Eignung für das Wissensmanagement, Master's thesis, Fachhochschul-Studiengang Informationsberufe, Eisenstadt, 2006, 31-32.

[Hunt and Mashhoudy, 2002] Neville Hunt and Housh Mashhoudy, Discovering Important Statistical Concepts Using SpreadSheets. Created at: January 29, 2002. Retrieved at: October 28, 2006. http://home.ched.coventry.ac.uk/Volume/vol0/philosop.htm.

[Smith, 2005] Waynes Smith, Graphics and Web Design Based on Edward Tufte's Principles. Created at: January 17, 2005. Retrieved at: October 29, 2006. http://www.washington.edu/computing/training/560/zz-tufte.html

[Tufte, 1999] Edward R. Tufte, The Visual Display of Quantitative Information. Created at: January 26, 1999. Retrieved at: October 29, 2006. http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html