Reading 12 and Assignment 12: Multi-Variate

by Mike Gleicher on April 1, 2017

Due: Initial Reading and Posting, Monday April 10th. Addition Postings by Friday, April 14. Discussions Close April 21.

For the readings you need to:

  1. Read #1 or #2 below (one of the old papers with ways to show multi-variate data) and get a sense of the range of things people have done.
  2. Try out parallel coordinates and scatter-plot matrices – since these are the two most important approaches
  3. Look at Colin Ware’s paper on bi-variate maps. Look at another paper on bi-variate maps if you want (it’s part of a discussion).
  4. Think about multi-variate glyphs. Read a paper if it helps.

Dealing with Multi-Variate Data is a huge topic – too much to cram into a single lecture. We’ve actually been talking about it all semester (but not explicitly). In fact, you might notice that there’s no Munzner reading for this week (since Multi-Variate is everywhere in the book), or Tufte (who talks about multi-variate all the time).

Look through one of these old papers are some of the (mostly bad) ideas for showing multi-variate data. You just need to skim through and get a sense of some of the techniques.

  1. – a gallery of different (old) ways to show multi-variate data
  2. Wong and Begerton. 30 years of Multidimensional Multivariate Visualization (find it here). 1997

The two most important methods (that survived the test of time) are Parallel Coordinates and Scatterplot Matrices (SPLOMS). There really aren’t readings. Play with some implementation or look at something to get a sense of their pros and cons. You don’t need to read a whole paper.

  1. parallel coordinates – find something that does it. There is a nice demo of a D3 implementation (here) that has a lot of nice features (play with the demo and see how selection and coloring makes parallel coordinates more palatable).
  2. scatterplot matrices – again, their all over. Check out the Scatter-dice system (project page). Rather than reading the paper, you can watch the video.

Another important problem is showing multiple fields in 2D. Just 2 is hard. Here is a nice method that is presented well. (again, don’t worry about the details, try to get the essence of the problem, alternative and solution):

  • Colin Ware, “Quantitative Texton Sequences for Legible Bivariate Maps,” IEEE Transactions on Visualization and Computer Graphics, vol. 15, no. 6, pp. 1523-1530, Nov./Dec. 2009, doi:10.1109/TVCG.2009.175  (ieee page, colin’s version)
  • In previous years, I had people look at other Bi-Variate map papers in order to appreciate how this approach compares/contrasts. You can look there (#1-3, and to a lesser degree 4 on this page).

For the design challenge, you need to think about multi-channel glyphs. These papers might help you, since they are extensive discussions of glyph designs. This is optional reading.

  • Taxonomy-Based Glyph Design, with a Case Study on Visualizing Workflows of Biological Experiments – (IEEE) – the problem is specialized, but the study of how to cram lots of information onto a glyph is interesting.

  • A Systematic Review of Experimental Studies on Data Glyphs – (IEEE, author) – mainly about how glyphs are evaluated.

For the Discussion (on canvas (HERE)):

  1. For the initial post: talk about the various solutions to the multi-variate visualization problem (the problem is diverse – the solutions you’ve read about are for a specific type). What might these various designs you’ve read about be good for? (in the old papers, and parallel coords and SPLOMS) Are there designs that you think are good/bad? (some of those old ones deserved to be forgotten, in my opinion)
  2. In a second post – talk about the bi-variate map problem. What are the pros and cons of Colin’s texton+color approach? What alternatives can you come up with?
  3. In a third post – the second part (phases 3 and 4) of Design Challenge 2 are clearly a multi-variate visualization problem. Talk about how the readings can help you come up with a good solution.

These 3 postings are the minimum. I encourage you to discuss the pros and cons of the different designs people bring up and to help each other think about Design Challenge 2.

Print Friendly, PDF & Email

Previous post:

Next post: