|
|
| Line 1: |
Line 1: |
| = Information Visualization, Task 3 =
| | http://www.clomidlmt.com/ buy clomid online wthout prescription - clomid no prescription |
| This article is the result of [[Teaching:TUW - UE InfoVis WS 2010/11 - Gruppe 01|group 1]] for [http://ieg.ifs.tuwien.ac.at/~gschwand/teaching/infovis_ue_ws10/infovis_ue_aufgabe3.html task 3] of the course [http://www.ifs.tuwien.ac.at/~silvia/wien/vu-infovis/ information visualisation] at [http://www.tuwien.ac.at Vienna University of Technology] in the winter semester 2010/2011.
| |
| | |
| The applications developed in the course of this task can be found on [http://infovis.trbuh.net http://infovis.trbuh.net].
| |
| | |
| = Data =
| |
| The data for this task is provided by the [http://www.wssinfo.org Joint Monitoring Program for Water Supply and Sanitation (JMP)] of the [http://www.who.int/en/ World Health Organization (WHO)] and the [http://www.unicef.org/ UN Children’s Fund (UNICEF)]. It contains information about access to (drinking) water and sanitary standards in countries all over the globe, intended to monitor the process towards the [http://mdgs.un.org/unsd/mdg/Default.aspx Millennium Development Goals (MDG)].
| |
| | |
| == Area of Application ==
| |
| This being a fairly ambitious program requires close analysis of the data. This being especially necessary as it is clear (and also stated so by the JMP) that the acquisition of such data is hindered by infrastructural (numerous measurement standards, difficulty to reach places for a significant sample, etc.) and financial shortcomings. Expectedly the samples are not complete. Several countries lack one or more variables either completely or partially (over sample period 1990-2008). To overcome these problems, solutions had to be found for the individual visualizations.
| |
| | |
| == Analysis of Dataset ==
| |
| The dataset provided is structured temporally (time periods in which the samples were taken) as well as ordinally (for measurement of water quality so called drinking-water and sanitation ladders have been defined [JMP, 2010]: piped water sources are better than other improved which again are better than unimproved sources) and it contains quantitative samples. Nominal values tell about the region a sample is related to. The samples themselves are absolute and/or relative figures. Thus the dataset at hand is multidimensional and temporally structured.
| |
| | |
| The given information was extended by the (nominal) information to which continent a country belongs. This implicit structure adds a hierarchy, as the countries can be summed up to continents – a fact that is necessary or at least favorable for some visualization techniques.
| |
| | |
| == Potential Users ==
| |
| By the nature of the data processed in this task, the potential audience is an extremly wide one. In order to serve a large part of it a multi-step approach was chosed for the presentation of the data. These steps target not only different interrest groups but also help to gain an overview as well as an understanding and feel for the details of the data and their implications for countries development towards the millenium goals.
| |
| | |
| Since not even academic audience usually has the same skill in reading and understanding graphs and data (e.g. a doctor and a statistician) this has been considered when creating the visualisation applications. This also explains why there is no ''best order'' in which to explore those apps. For example might a statistician first choose to view the scatterplot and by this gain an overview of the distribution of data and correlations within the sample. Development workers on the other hand will likely be interrested in absolute numbers of people that lack access to improved water to decide where to deploy missions for improvement. And a politian or decision maker could want to have relative figures to base her or his decission for further support programms on them.
| |
| | |
| For this reason a webpage has been developed that gives acces to the three developed applications (including a brief description) and as such makes it possible for a broad audience to view the data they are (personally or professionally) interrested in.
| |
| | |
| = Objectives =
| |
| According to [JMP, 2010] a main goal is to identify disparities in development of different regions, so well-directed actions can be taken. The visualization shall support to examine:
| |
| * Which regions are left behind?
| |
| * Differences between urban and rural regions
| |
| * Differences between provided services (water/sanitation)
| |
| * Performance of different regions (with respect to different baselines where they started from)
| |
| | |
| = Visualization =
| |
| | |
| In the visualization solutions supplied, several different techniques are utilized to present the data:
| |
| * Bar chart
| |
| * Line chart
| |
| * Stacked time series
| |
| * Scatterplots
| |
| | |
| Based upon examples of the Protovis toolkit, interactive charts have been enhanced, adopted or interlinked in order to become more powerful (access them at [http://infovis.trbuh.net http://infovis.trbuh.net]). The driving idea behind this was Ben Shneidermans information seeking matra [Shneiderman, 1996]: "Overview first, zoom and filter, then details-on-deman"
| |
| | |
| == Development of Drinking Water Supply ==
| |
| | |
| === Visual Mapping ===
| |
| | |
| ==== Bar Chart ====
| |
| [[Image:Watervisbarchart.png | 350px | thumb | '''Figure 1''' : ''Bar chart'']]
| |
| In bar charts the length of the respective bar expresses the corresponding value of the variable. The advantage of this technique is the ability to feature individual values and support comparison of one value to another [Few, 2007, p. 3]. In the given example in figure 1 a ranking of the improvement/decline of improved drinking water access in the time period between 1990 and 2008 has been made.
| |
| | |
| In addition a bar can easily be used to express a second variable by color. This complies with Tufte´s design principle of multifunctional graphical elements [Tufte, 1983, p. 138]: “Mobilize every graphical element, perhaps several times over, to show the data”. Color is used to show the increase/decrease of population in the corresponding country. This technique makes it possible for the user to see correlations between variables (in this case improvement of drinking water and demographics) as well as differences between different groups (e.g. urban and rural).
| |
| | |
| ==== Line Chart ====
| |
| Line chart are suitable to show development, and subsequently change, over time. When using bar charts, the overall shape of change gets lost in a forest of bars [Few, 2007, p.3]. Thus line charts are used to display evolution over the years.
| |
| | |
| === Interaction ===
| |
| | |
| Following interaction techniques have been applied for this purpose:
| |
| * Overview and details on demand
| |
| * Filter
| |
| | |
| ===== Overview and Detail on Demand =====
| |
| Overview and detail shows the data at more than one level, but they also show where the finer grain display fits into the larger grain display. Their main disadvantage is that they require coordination of two visual domains [Card, 1996]. In this visualizations overview is used to provide the big picture, displaying all countries. When pointing with the mouse at an element, the corresponding detail data is shown in a different view.
| |
| | |
| ===== Filter =====
| |
| With filter functions the user can select different subsets of the whole dataset. By allowing users to control the content of the display, users can quickly focus on their interests by eliminating unwanted items [Shneiderman, 1996]. In the visualization the user can filter regions and parts of the population (urban/rural), by doing this interesting aspects of the dataset can be found (for example improvements in rural South America).
| |
| | |
| | |
| == Proportion of Regions with Access to Unimproved Water Sources ==
| |
| [[Image:Watervis-stackedtimeseries.jpg | 350px | thumb | '''Figure 2''' : ''Stacked time series'']]
| |
| As pointed out above, the focus was set upon development of regions over time. In this application visualization was restricted to utilization of unimproved water sources and its development. Of course this can be easily extended onto other variables. Important aspects are: how do different regions develop, and which regions are concerned most?
| |
| | |
| === Visual Mapping ===
| |
| | |
| ==== Hierarchical Stacked Time Series ====
| |
| | |
| According to [Few, 2006] lines can be used for time series to emphasize overall patterns, with time being placed on the horizontal axis.
| |
| | |
| The vertical axis is used to show how regions contribute to the total number of people with access to unimproved water sources. Instead of lines filled areas are used to depict the numbers, by this multiple countries can be stacked (adding up to the total number of each time period). The region taking up the most part of the chart is the one with the highest impact.
| |
| | |
| Color is used to encode the type of the region (urban, rural). And a text-label is used to identify the region.
| |
| As suggested in [Ward et al. 2010], redundant mapping is used to emphasize important data: regions with higher impact. For this labels are scaled proportionally, and color-saturation expresses the influence of the country/region on the total number.
| |
| | |
| ==== Bar Chart ====
| |
| So far this chart does not show if the number of a region actually decreases or increases, as the numbers are compared to the total number per year. Thus it is linked to a bar chart. This chart again places Time on the horizontal axis, and quantity on the vertical –although absolute figures. The bars are used to display the figures of the country/region the mouse hovers over. For comparison reasons each year’s global total is also displayed.
| |
| | |
| === Interaction ===
| |
| At any point the user can filter the shown data in the following ways:
| |
| * Choose if rural, urban, or total data of the countries/regions is shown
| |
| * Filter the countries/regions by their name, via search query
| |
| | |
| The first view shows the development of continents. The user can inspect how countries contribute to a continents’ development by clicking on the continent (zooming in). This brings up the name of the selected continent in the top left corner, providing basic contextual information.
| |
| | |
| When filtering and zooming, the scale is adjusted to the amount of regions presented in order to make regions with small impact visible as well. Moving the mouse over a region selects it for comparison to the total number in the bar chart on the right. This also clarifies the impact of this region to the total number.
| |
| | |
| == Scatterplot ==
| |
| [[Image:ScatterplotKeyvariables.png | 350px | thumb | '''Figure 3''' : ''Scatterplot'']]
| |
| === Data Mapping ===
| |
| According to [Mazza, 2009, Table 3.2, p.40] there are three suitable preattentive attributes for the visualization of quantitative data: length and numerosity (form) as well as a spatial position in 2D-space.
| |
| | |
| Scatterplots, the third representation chosen for the given quantitative data, are based on this last attribute – 2D spatial positioning. Further this “standard scatterplot” is enhanced by utilization of color intensity. This attribute, suitable for (preattentive) visualization of ordinal data [Mazza, 2009, Table 3.2, p.40] is used to visualize the density of samples at a certain position, which translates into an ordinal relation. This additional feature uncovers what would otherwise remain hidden within the amount of data: aggregations of same or similar samples. This also "reduces" the space needed per plot which is one of the shortcomings of scatterplots – especially when plotting multiple variables.
| |
| | |
| === Interaction Features ===
| |
| According to [Stuart Card, Informaiton Visualisation, p.525] scatterplots are a sensible visualization technique for problems with one or more variables. In the latter case they are usually combine to scatterplot matrices. First the user has to choose the data to be displayed by the scatterplots, which can be either a full-variable display of the water markers or sanitary markers. The third set consists of the total figures of water and sanitation. Each of these three sub-sets also contains the development of the countries’ overall populations and the development of urbanization.
| |
| | |
| Via brushing with the mouse the user can then select areas/countries within any of the scatterplots. The chosen selection is then highlighted in all the other scatterplots as well. With this a general overview of the data is possible and eventual correlations between variables can be uncovered.
| |
| | |
| | |
| | |
| | |
| == External Links ==
| |
| * [http://www.wssinfo.org/data-estimates/table/ JMP-download of raw data]
| |
| * [http://vis.stanford.edu/protovis/ Protovis homepage]
| |
| * [http://infovis.trbuh.net Developed visualization applications]
| |
| | |
| == References ==
| |
| * [Card, 1996] Card, Stuart: "Information visualization and information foraging", Proceedings of the workshop on Advanced visual interfaces - AVI ’96 (1996), 12. URL: [http://portal.acm.org/citation.cfm?doid=948449.948451 http://portal.acm.org/citation.cfm?doid=948449.948451], last access on January 17, 2011.
| |
| | |
| * [Card, 2008] Card, Stuart: “Information Visualization”, in The Human-Computer Interaction Handbook: Fundamentals, Evolving Technologies, and Emerging Applications, Sears, A. and Jacko, J., A. (eds.), Lawrence Erlbaum Assoc Inc, 2008
| |
| | |
| * [Few, 2006] Few, Stephen: "Table and Graph Design at a Glance". URL: [http://ieg.ifs.tuwien.ac.at/%7Egschwand/teaching/infovis_ue_ws10/download/Table-and-Graph-Design-at-a-Glance.pdf http://ieg.ifs.tuwien.ac.at/%7Egschwand/teaching/infovis_ue_ws10/download/Table-and-Graph-Design-at-a-Glance.pdf], last access on January 17, 2011.
| |
| | |
| * [Few, 2007] Few, Stephen: "Visualizing Change", Visual Business Intelligence Newsletter (2007), p.1-15. URL: [http://www.perceptualedge.com/articles/visual_business_intelligence/visualizing_change.pdf http://www.perceptualedge.com/articles/visual_business_intelligence/visualizing_change.pdf], last access on January 17, 2011.
| |
| | |
| * [JMP, 2010] Joint Monitoring Programme: "Progress on sanitation and drinking water – 2010 Update", (2010) WHO Library Cataloguing-in-Publication Data. URL: [http://www.wssinfo.org/fileadmin/user_upload/resources/1278061137-JMP_report_2010_en.pdf http://www.wssinfo.org/fileadmin/user_upload/resources/1278061137-JMP_report_2010_en.pdf], last access on January 17, 2011.
| |
| | |
| * [Mazza, 2009] Mazza, Riccardo: "Introduction to Information Visualization", Springer-Verlag, London, 2009
| |
| | |
| * [Shneiderman, 1996] Shneiderman, B: "The eyes have it: a task by data type taxonomy for information visualizations", Proceedings 1996 IEEE Symposium on Visual Languages (1996), p.336-343. URL: [http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=545307 http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=545307], last access on January 17, 2011.
| |
| | |
| * [Tufte, 1983] Tufte, E.R.: "The Visual Display of Quantitative Information", Cheshire, Connecticut, Graphics Press, 1983.
| |
| | |
| * [Ward et al., 2010] Ward, M., Grindstein, G. and Keim, D.: "Interactive Data Visualization: Foundations, Techniques, and Application", A K Peters, 2010.
| |