Visualizing with Text releases any day now: I hope to have my copy before the end of VisWeek. I’ve finally posted all the figures that I authored on-line with a CC-BY-SA 4.0 license. There’s 158 high-resolution images and diagrams from the book in the PDF file. These may be a nice complement to the eBook or physical book as some of the text may be too small to be readable in some of the larger visualizations. My figures are all released with a CC-BY-SA license, so they can be reused, for example, for teaching or mixed up into a collage or whatever.
There’s another 99 figures that are not mine – I’ve included links to online versions of these images where available on the last page.
Sometimes people ask me which of the visualizations I like the best. The answer varies over time, although I am currently biased towards the text-dense multi-variate visualizations designed for a large screen, such as these ones (Figures 6.19, 8.10, 9.8, 10.11, 11.3, 12.11) – see the PDF for high-res versions:
Why? Viscerally, I like the rich texture of shapes, colors, and structure where multiple patterns appears – visualization should be supportive of representing complexity and affording multiple interpretations. In my day-to-day work, I often design visualizations for financial market professionals: they don’t necessarily make money if they have the same ideas and same thesis as everyone else. Data-dense visualizations that prompt multiple hypotheses can be a good thing. (see also Barbara Tversky’s keynote at Visualization Psychology earlier today!).
I also think these dense visualizations push the boundaries of the design space of visualization and of text-visualization. Perceptually, multi-variate data can be a challenge. Data-dense visualizations can be a challenge. The linearity of text (i.e. you have to read words in some order) vs. the volume of information is a challenge: what happens to the global pattern? what happens if “overview first” doesn’t necessarily provide a macro-pattern?
A couple of these visualizations I just presented for the first time yesterday at the Visualization for Digital Humanities workshop in a paper titled Literal Encoding: Text is a first-class data encoding.