Note that I've assigned an additional reading to help you with this assignment: Kosslyn, S. M. (1989) Understanding Charts and Graphs. Applied Cognitive Psychology, 3. 1989, pp. 185-226.

Optional Supplemental Video: How to Display Data Badly, lecture by Howard Wainer

(A) Mixing and Matching Data Types.

Choose a domain (e.g., web server logs, network performance, consumer buying patterns). Describe an appropriate graph, chart, table, diagram, etc in that domain for each of these combinations of input type:

nominal x nominal
nominal x ordinal
ordinal x ordinal
ordinal x interval
interval x interval

You can use visual displays you find produced by someone else using real data, or sketch out what such a visual display should look like assuming you had real data (as done in class for the web page access examples).

(B) Critiquing visualizations.

Read sections 1, 3, and 4 of the following article describing experiments in evaluating web page access patterns:

     Tauscher, L. and Greenberg, S. (1997) How People Revisit Web Pages: Empirical Findings and Implications for the Design of History Systems. International Journal of Human Computer Studies Special issue on World Wide Web Usability, 47(1), p97-138, Academic Press.
    

Use principles and observations from Tufte, Kosslyn, and class lecture, as well as your own judgement, to assess and critique figures in this paper.

1. Consider Table 2.  Is there anything potentially misleading or confusing about the way it is laid out?
2. Assess the portrayal of information in figures 5a and 5b.  What are the authors trying to convey with these? Would some other kind of graph be more appropriate for this? If so, which?
3. Critique figures 6a and 6b in terms of graphic display principles.  Draw examples of better, perhaps non-standard, ways to display the same information, and perhaps additional information as well.
4. Figures 7 and 8 attempt to show some non-standard kinds of information.  How well do they succeed at their task?

(C) Critique or Create.

Do one of the following two options:

1. Design a better way to do the codings for the recipe book that Barbara Rosario showed us in class.

   Unfortunately the book seems to have been cleaned out along with the rest of the PhD room, but below is an attempt to reconstruct the essential elements. It won't be possible to compare to the original, so instead just justify your choices.

   The book used primarily stacked color coding to display:

   1. the total calories of the recipe (and the calories from carbohydrates, fat and fibers).
   2. the proportions of the calories with regard to the average daily intake (2000 calories)
   3. a qualitative measure of the recipe calories (hyper - hypo -medium caloric)
   4. qualitative measures for sodium, sugar and fiber (high, medium, low)

2. Critique this visualization from Wired Magazine.

(D) Start Planning.

Think about what dataset you'd like to use for the next assignment. You will be using several tools for exploring multidimensional datasets. These tools use brushing and linking among other techniques.