Group items tagged

It would be interesting to pull the data in from previous decades and see how the elevation has changed in different areas.

@adrian – it’s just raw totals, grouped geographically. These aren’t scientific by any means, I basically took the underlying pattern and extruded it out and smoothed it a bit to make it look “pretty”. But basically each image is the aggregate numbers for a single year of crime data.

@richard – yes, there is some smoothing in effect, which means that the ridge along Shotwell St (for the prostitution map) is indeed a bit smoothed between peaks. That’s not to say that there are only two peaks at Shotwell and 19th and Shotwell and 17th. There are incidents in between as well, but the big peaks at those major intersections does mean that the ridge between them appears higher than the actual incidents along those blocks support. A lot of people have commented on the usefulness of maps like these. I want to stress once again: this was done as an art project much more than a useful visualization. My goal was not to provide useful information that one could act on.

“one trick pony. these maps add nothing of value to a standard color plot.” I disagree: allowing for a third dimension of elevation makes the reality of concentration clearer – and half the point of crime mapping is to measure concentration, not simply “intensity.”

Great idea and nice work on the graphics, but there are at least three improvements you should make to reveal *true* patterns. Forgive me if you already did these. 1) Availability bias – normalize for population density (i.e. per capita activity) 2) Sampling bias – normalize for the number of cops on the beat (geographic and crime type) 2) Frame bias – break it up by daytime and night time