Inspired by Robert Hodgin‘s impressive work, I decided I wanted to take a crack at making animations in Processing. Inspired by Rob Goodspeed‘s thoughtful experiments, I figured I might as well try to make it about something interesting.
The DC Office of the CTO makes a lot of interesting data available, so I decided to take a crack at doing something with it. Here’s my first pass: it’s a visualization of every reported crime incident in 2007.
Blue means property crime — robbery, larceny, etc. Red means violence — assaults, murders and the like. Green means sex crimes. The color-coding is obviously less than perfect, but is at least sort of interesting. Each frame in the video represents 8 hours — one police shift. Colors gradually fade to a dark, translucent gray after two weeks’ worth of time has elapsed. You can view a larger version of this video by clicking through to the movie’s Vimeo page, and an even better-quality version by then scrolling down to the lower right for a Quicktime download link. You’ll need to log in or create a Vimeo account for the Quicktime — but you should! Vimeo is the best videosharing site I’ve used by a mile.
At the moment I’m trying to figure out where I should go from this point. A few ideas:
- Add graphs and a calendar display to help users see the relative levels of crime over the course of the movie
- Identify and place a street map under the crime visualization
- Produce a static, high-resolution render showing the differences in location of crime between 2007 and 2008 (YTD)
- Do it all over again with the city’s pothole data (also geocoded!)
- Add a jaunty soundtrack
If you have any other suggestions, please let me know. Those of you interested in the geeky details should keep an eye on EchoDitto Labs — if I can find time I’ll be posting an accounting of the tools I used over there.
City-mandated disclaimer after the jump
The data made available here has been modified for use from its original source, which is the Government of the District of Columbia. Neither the District of Columbia Government nor the Office of the Chief Technology Officer (OCTO) makes any claims as to the completeness, accuracy or content of any data contained in this application; makes any representation of any kind, including, but not limited to, warranty of the accuracy or fitness for a particular use; nor are any such warranties to be implied or inferred with respect to the information or data furnished herein. The data is subject to change as modifications and updates are complete. It is understood that the information contained in the dataset is being used at one’s own risk
I think it would be interesting if you could aggregate the data a bit. For example, instead of a dot for every crime and a from every shift, would it be feasible to aggregate it into geographic and temporal buckets (with each bucket representing say a square mile and a week) and then display a dot whose size or color indicates the total number of crimes during that period?
Another interesting extension to this would be to adjust for population, so that the dots represent per capita crime rates. Right now, it looks like most of the crime happens in a cluster slightly North and West of the capital. I’m assuming that that’s largely because NW is where a lot of the people are.
Naturally, I don’t know how much work any of these ideas would be.
Those are great suggestions, Tim. The one about crime per capita had occurred to me, too — I’d also like to plot it versus income. But I think I’m going to have to go pester Mr. Goodspeed about how to go about getting that data.
The aggregation idea is one I can pursue immediately, though.
Why did you fade the colors to gray? I’d like to see the cumulative differences in distribution of the three categories of crime. Fading to a lower intensity of the original color would allow that.
The reason I faded them to gray was so that “current” crimes would be emphasized — even if older crimes are darker, it’s hard to pick out currently-occurring crimes over top of them.
But I agree that that would be an interesting visualization, and I have tried running it that way. I think it will be just as good as a high-quality cumulative render, though. I’ll try to find some time to do that soon.