Brain Lint

A couple of weeks ago I had the good fortune of sitting in on lecture of a scientific visualization class* at Tufts at which Ben Fry, creator of many great works of visualizations that can even be called information art as well as the visualization toolkit, Processing, was guest speaking. The talk was great, spanning lots of work and interesting commentary.

Some notes:

• Ben showed quite a bit of his previous work – some of it would be familiar to readers of his book, Visualizing Data.


• Showed off some of his work that has appeared in movies, highlighting the fact that he is asked to add rows of standard grey computer buttons to his work because it doesn’t look “real” otherwise.


• Talked about some experience teaching classes, particularly the challenges of classes with mixtures of cs students and artists. Making CS students do projects more artsy and artists do more interactive, technical work can be interesting. He showed off some examples of student work. (One cool student project asked a set of Nobel laureates what type of pets they had. Quite a few found time to respond and the results are here.)

• The coolest demos were of some of the work he’d done for Oblong Industries (Not a lot of information online now- here’s one cnet article)- they have a working Minority Report-style gesture interface that allows one to control a computer with hand movements. Paired with the right interface, this looks to make light work of navigating through vast amounts of multidimensional data. Ben showed some videos, along with a demo (running on his macbook pro w/o the fancy hardware it was still really cool).

• I’d been asking for a class like this to be offered several times while I as still working on my degree at Tufts, but to no avail. Of course it’s offered right after I graduate!

I love this interactive graph showing the margin of victory in each of the primary states for Clinton and Obama. It blends a histogram with labeled blocks to really effectively illustrate the data. The animation between different views of the data puts it over the top.

I noticed this morning that google finance has a new stock screener feature that lets you choose stocks with features in a certain range by way of an interactive sparkline. These are miniature graphs that go inline with text. In this case the graph is a histogram that indicates how much of the stock market falls into each part of the range – this will give one a quick preview how inclusive their search parameters are.

Oil change international put together a great tool to visualize the flow of money from oil companies to presidential candidates and congressional representatives. The graph view of the presidential race is sort of what you expect, with republicans soaking up more oil money than democrats.

What interests me about the data though, is what makes a donation from someone who works at an oil company “oil money”? Where do we draw that line? It would seem that a matched pair of max $4600 donations from the CEO and (homemaker) spouse are on one end of the special-interested donation spectrum, but what of a $500 donation from someone who owns a gas station, or a pair of $500 donations from a research scientist?

For me, I think the inclusion of some of the donations as oil-money are disingenuous, but its hard to say which donations are or are not to be included.

Maybe one day we’ll see public campaign finance and no one will have to figure that out?

In the image below I’ve plotted the average gas price in each state for 4/25/07 (data from here) with and without state per-gallon taxes included. Without the taxes included, it becomes obvious that gas prices increase on the west coast, perhaps due to transportation costs? ( a quick search didn’t turn up any port-by-port oil import stats).

I created this using ruby-shapelib and rmagick as mentioned previously.

I found a document (pdf) on the US Census Bureau site with data on the mean and median population centers of the United States over time. Unfortunately they didn’t stick it on a map so one can get a real sense of the slow south west migration. Fortunately this provided me with a good excuse to learn the google maps API and put the points on a map.

I’ve put the maps up here. I’m not sure I expected the median and mean to be so similar (though unfortunately there’s a hundred years less data for median). The mean ends up further west than the median because the distance of the west coast amplifies the weight of its population.

I was actually looking around the US Census site looking for some interesting data to build a Treemap with, but didn’t find anything yet with a two level hierarchy that didn’t end up at the county level. Got any ideas of something that could be interesting in treemap form?