I have been absent for too long — this time for a good cause: Vacation! Geekgirl enjoyed California and Vegas and anything not involving a computer for over a week. It is a good experience to have at least once a year. And in return, dear reader, I give you a nice long post. Thank you for waiting for me!
I find data visualization fascinating. A good visual can more fully tell you what your data is “about” than any string of numbers. Take, for example, the cool interface at the new travel site Hipmunk.
A simple interface modification changes a list of prices and times into a much easier to process visualization of the data — flight times turn into horizontal placement, airlines into colors, and layovers are also clearly marked. You can sort by price, time, or “agony” — their own combination that calculates both the pain to the pocketbook and the pain of a horrible red-eye flight with a 5 hour layover in Chicago. No, I don’t work for Hipmunk, I just think this is really great data visualization!
Make your own
And if you want to make your own data visualizations, there’s a really cool tool now for free — Tableau Public — which will let you create some pretty stunning, interactive data on your own. The only bummer is that it’s windows only, so I couldn’t play with it myself. But here’s a screenshot of an interactive data set of Japanese earthquakes since the 1900’s (you can click the link to actually experience the interactivity):
Dance the Algorithm
And for those of you who are computer geeks, I was sent this human-enacted visualization of computer algorithms. This same series has a variety of different algorithms depicted as dance, including shell-sort and select-sort. I don’t know computer science from my ass, and was a bit disappointed that I didn’t actually learn anything from the video, but I guess that this is one of those cases where some prior knowledge goes a long way.
Learn through visualizations
This isn’t really data visualization, but it fits in this post because I think this gets at why images and visuals are so helpful to us as humans. The pictoral representation of a concept or idea can go further than just accurately representing the concept or idea — it can tell us how to think about it. The way data is presented helps us to interpret the numbers. Similarly, how we depict scientific ideas can help us to understand them. In that vein, Felice Frankel (who does wonderful projects on scientific images as art) recently embarked on a project called Picturing to Learn. College students were told to draw a picture to explain a particular scientific concept to a high school senior. The results were meticulously documented on the website, and give insights into how students think about different topics. You can register for full access to the database, and learn to interpret student drawings for yourself. A great tool for teachers!
For non-beautiful charts, see JunkCharts, which publishes offending graphics full of chart junk (a term coined by Edward Tufte, to describe the gratuitous use of graphics and other visual elements that don’t help the viewer to interpret the graph or, worse, obfuscate the information therein).
And some final humor
And of course, any post about cool data and charts wouldn’t be complete without my favorite tongue-in-cheek pie chart (which I found on FlowingData, but thought was originally on GraphJam — see the comments on the FlowingData post for some attribution as to its origins):