Today I watched an old TED Talk by Dr. Hans Rosling entitled "Debunking third-world myths with the best stats you've ever seen." If you haven't seen it yet, I recommend taking a moment to watch it - I've never seen statistics presented in such an engaging and entertaining fashion. In this talk, Dr. Rosling uses his fantastic visualization software to demonstrate the changing relation between the wealth and health of nations over several decades.
Today I watched an old TED Talk by Dr. Hans Rosling entitled "Debunking third-world myths with the best stats you've ever seen." If you haven't seen it yet, I recommend taking a moment to watch it - I've never seen statistics presented in such an engaging and entertaining fashion. In this talk, Dr. Rosling uses his fantastic visualization software to demonstrate the changing relation between the wealth and health of nations over several decades. The graphs percolate, flow, and dance across the screen with Rosling right there with them as he narrates.
With the right visualizations, data is exciting!
My favorite visualization tool is Panopticode, which generates treemaps of a project's cyclomatic complexity and test code coverage. You get so much information from that picture - you can see which areas of the code are problems. For me, a big black blotch has a much more emotional impact than a relatively large number.
We used the Panopticode cyclomatic complexity graphs as a big, visible chart on a pilot Agile project. The graph gathered a lot of attention; passers-by would stop and stare, then ask what the picture meant. It was a great conversation starter - we got to speak about refactoring, continuous integration, and simple code to everyone from software developers to testers and business users. Our big win was getting to speak with the product owner; he didn't understand code, but he understood the scope and the implications of code that was complicated and untested. There's no way a printout of a table of numbers would have gotten this attention.
So what do statistics on worldwide health and wealth have in common with cyclomatic complexity? They are both better understood with the right graphs.
I've always wanted to collect the sequence of graphs generated by Panopticode and turn them into an animation, the code equivalent of the YouTube video "Girl takes pic of herself every day for three years." I wonder if viewing the graphs over time would provide meaningful information about the project. For a team that practiced frequent refactoring, breaking large classes into smaller ones, would the movie look like Conway's Game of Life?
What's your favorite data visualization tool? How has it helped you?
