All 5 books, Edward Tufte paperback $180
All 5 clothbound books, autographed by ET $280
Visual Display of Quantitative Information
Envisioning Information
Visual Explanations
Beautiful Evidence
Seeing With Fresh Eyes
catalog + shopping cart
Edward Tufte e-books
Immediate download to any computer:
Visual and Statistical Thinking $5
The Cognitive Style of Powerpoint $5
Seeing Around + Feynman Diagrams $5
Data Analysis for Politics and Policy $9
catalog + shopping cart
New ET Book
Seeing with Fresh Eyes:
Meaning, Space, Data, Truth
catalog + shopping cart
Analyzing/Presenting Data/Information
All 5 books + 4-hour ET online video course, keyed to the 5 books.
Driving distance tables

I've been asked to show stats for pairs of items, where every item is paired with every other item. The typical way to do this is with a table that has the list in a set of rows and columns, and the statistic at the intersection, like driving distance tables:

bostonnew yorkwashington
bostonna200400
new york200na230
washington400230na

I've found a couple of interesting graphical representations that might be useful if annotated. The first is a link diagram and the other is a, well, I don't know what to call it, but it's interesting. Line thickness, color and annotation could be used to convey some metrics.

Does anyone know of examples of similar charts? I'm curious about ways that might make it easier to visualize the distance, show multiple statistics per pair, or otherwise highlight information between pairs of items that are interconnected.

-- Mike Combs (email)


Ooops. Somehow, while editing that HTML table by hand, I cut out the link to these New Zealand Driving Distance charts. They reduce the chart size by eliminating the redundance; Boston-New York distance is the same as New York-Boston. They're a bit harder to read, especially the first example, but the second chart is pretty clear, IMHO.

-- Mike Combs (email)


link diaram layout

An interesting interactive display that solves the link diagram layout problem with animation is at NetScan. (I found this following links from the UMD TreeMaps history discussion!)

-- William RIcker (email)


Diagrams similar to the "what to call it" are traditional in biology to show the reproductive relationships (i.e. fertility or sterility upon crossing) among different species and varieties of plants. The names of the plants, each enclosed in a small circle, are arranged in a larger circle; lines going from one small circle to another encode the relationship beteeen the two joined species. For example, a thin line connecting the species might indicate they are intersterile, a thick line that they are completely interfertile, an intermediate line that crosses between them are partially fertile. The diagrams often contain additional information, either as annotations or symbolically (e.g. sample sizes of experimental crosses, geographic distribution, morphology, etc.). In some diagrams, the species are arranged irregularly (i.e. not in a circle).

Unfortunately, I don't know of any examples on the web; a number can be found in Verne Grant's Plant Speciation (Columbia University Press, 1971).

Such diagrams are used less often than they used to be. A paper in the latest issue of Evolution (vol. 57, no. 6, June 2003) has a dataset of just this kind (on birds, rather than plants, though), but the data are presented in a table, rather than a diagram (there are various good reasons why the authors may have chosen a table over a diagram).

-- Gregory C. Mayer (email)


I've now found some plant fertility diagrams. The following is a circular one: http://www.biologie.uni-hamburg.de/b-online/e38/8.htm while this one is non-circular: http://www.biologie.uni-hamburg.de/b-online/e38/7.htm Both are redrawn by Peter v. Sengbusch at a very nice website: http://www.biologie.uni-hamburg.de/b-online/e38/38c.htm

-- Gregory C. Mayer (email)


Great stuff, especially the plant diagram in three layers, because it lends itself to automated creation and printed output. Thanks!

-- Mike Combs (email)


While this isn't the ultimate solution, I thought I'd post the mileage chart concept. In this example, latency between NYC-BOSTON is not the same as BOSTON-NYC, so I put two values in each cell.

This format provides two advantages over the matrix above. First, column labels can be longer without stretching the column sizes (thus allowing more columns). Second, it only takes one lookup to see data for both directions. Reading the data for one city is easier as well. I'm not sure, though, if the arrow icons do the trick in explaining the latency, particularly since they kind of change direction from row to column. Also, this doesn't allow any sorting or grouping of the data, as the suggestions other people posted might.

In implementation, I'd probably highlight the row and column as the mouse passes over the city name.

Latency
Boston (2.242)
→37
←42
New York (12.115)
→37
←42
→37
←42
Atlanta (93.5)
→37
←42
→37
←42
→37
←42
Dallas (51.201)
→37
←42
→37
←42
→37
←42
→37
←42
Houston (11.57)
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
Baltimore (37.15)
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
Washington (191.25)
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
St. Louis (97.210)
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
Phoenix (53.12)
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
Albequerque (39.27)
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
→37
←42
Seattle (93.163)

-- Mike Combs (email)




Threads relevant to design case studies:


Privacy Policy