10 comments

[ 4.0 ms ] story [ 24.1 ms ] thread
I'm not too sure about the visualization they're using; I see red in the heatmap for San Francisco, but Austin and Englewood on my Chicago map are yellow/green --- nothing is red. There is no neighborhood in San Francisco that has, for instance, a recurrant problem with dog fighting, as these Chicago neighborhoods do.

I wonder if they're averaging something out over the city and diluting the scores that way.

Yeah, they're not directly comparable, unfortunately, because the police departments in the different areas have different reporting habits, and this is all based on frequency of police reports.

EDIT: Within single police jurisdictions one area should be comparable to another, though.

Hi sir, big fan here.

Was just wondering when you'll be coming out with a mapper for everything else on craigslist ?

Is dogfighting the kind of crime people are worried about though? I suppose it's an indicator that a neighborhood is home to undesirable residents, but I'd imagine most people are more worried about theft, vandalism, and violent crime.

It's not integrated with Craigslist but this site is pretty good at showing details: http://www.crimemapping.com/map/ca/sanfrancisco

Funny story, the first time I looked up my address on there I was horrified by the number of "public urination" citations - it was ridiculously high even by San Francisco standards. Then I remembered that Bay to Breakers had gone by my apartment 3 days previously.

Yes, I'm using it as a synecdoche for "intense gang violence".
It would be even more informative if the map could be adjusted for population density.

Compare two neighborhoods: Neighborhood A has a population of 100 and crime rate of 1 per month. Neighborhood B has a population of 10000 and crime rate of 1 per week.

Assuming the crimes are equal, in neighborhood B you are ~4% as likely to be a victim of a crime as you would in the smaller neighborhood A. But in the current map, neighborhood B would look like it has a bigger crime problem.

Hm... What I really need is foot traffic - Commercial areas have a huge amount of foot traffic but very low population density.
Foot traffic would be great for places like suburban malls or the weekends, but is less relevant for things like burglaries and car thefts (unless its burglaries at businesses). For city centers, worker population density would work just as well during the week, since workers are going to be the most likely to be in an area from 8am-6pm.
Anyone know where you can get data to feed something like this? I've been wanting to do a visualization for my current city, but don't know where to go about getting the info and our police department is far from technical...
Yep, as mentioned in the article, the data is from SpotCrime.