November 12, 2005

crime statistics

Filed under: — cwage @ 11:12 am

I was inspired by the discussion on this post to do a little digging on the Nashville crime situation. I have often hear people talk about Inglewood and east Nashville (I know better now than to conflate the two) as if they were one step away from Compton, crime-wise. People also seemed to have this impression about Melrose (37204), where I used to live, when in reality it was a pretty nice, if run-down neighborhood in the unfortunate but inevitable throes of gentrification. There was crime, occasionally, of course, but not to the extent that I was afraid to go for a walk at any given time.

So I was curious where this impression came from -- what are the "worst" neighborhoods in Nashville? Turns out, defining "worst" is tough.

Nashville's police department has some numbers which are interesting. According to these numbers, 37211 (south nashville), 37013 (antioch), 37206 (east nashville), etc. had the highest total criminal incidents. But these numbers don't really tell the whole story. These areas could have more crime because they simply have more people. A statistic that helps clarify this is the percentage increase in crime, which is helpful to pinpoint demographic shifts as well as areas having problems, but it still doesn't answer the real question I have, which is:

While living in a given area of Nashville, how likely are you to directly experience a criminal act? The ideal answer to this is really crimes-per-capita. Unfortunately there are a few problems in attempting to calculate this:

First, the most recent population by zipcode data available is via the 2000 Census's ZCTA (Zip Code Tabulation Areas). The population of Nashville hasn't changed too drastically between 2000 and 2005, however the demographics could (and probably) have shifted between areas in those five years.

Second, on a small scale like this (metro area), sample set is a problem. Naturally if you have a zip code with 2 people living in it, and there's one crime, suddenly you have an 0.50 per-capita crime rate.

Nonetheless, I forged ahead, and here's what I found:

Zipcode Population Crimes Crimes per capita
37086 18708.00 1 0.0001
37015 16600.00 11 0.0007
37027 35747.00 27 0.0008
37143 4333.00 5 0.0012
37080 7394.00 52 0.0070
37221 34809.00 467 0.0134
37138 22160.00 367 0.0166
37215 22112.00 398 0.0180
37220 6163.00 113 0.0183
37076 29547.00 604 0.0204
37072 26386.00 567 0.0215
37212 18547.00 446 0.0240
37205 21861.00 553 0.0253
37218 15079.00 480 0.0318
37013 51343.00 1758 0.0342
37211 64753.00 2258 0.0349
37217 28293.00 991 0.0350
37216 19132.00 712 0.0372
37209 35081.00 1416 0.0404
37214 26474.00 1107 0.0418
37115 35192.00 1568 0.0446
37204 11024.00 515 0.0467
37189 2343.00 115 0.0491
37207 35744.00 1987 0.0556
37206 27751.00 1631 0.0588
37210 16170.00 1039 0.0643
37208 15272.00 1095 0.0717
37203 12781.00 1459 0.1142
37219 830.00 194 0.2337
37228 331.00 111 0.3353
37201 1167.00 421 0.3608
37213 137.00 136 0.9927

Here you can obviously see the sampleset problems I was talking about. My zip code, 37201, doesn't look so hot. 37213 has a nearly 1.0 crime-per-capita rate -- 1 crime for every person! Of course, that's because it's the area around the stadium and, like downtown, more people come here than actually live here. They just come here, watch football and/or get drunk, and then commit drunken crimes. Good times!

So there you have it. There are no real shockers here. The wealthier suburbs are relatively crime-free, and the downtown area is prone to crime even in areas where there really aren't any residents. But it is interesting to note that by these numbers, East Nashville is only marginally less safe than, say, Green Hills.

Sidenote: in the course of searching for maps of zip codes, I found this really cool site which interactively maps zipcodes by digit. Give it a try.

  • t. rev

    It might be useful to compute crime rates based on a pro-rated population figure, say 50% resident population + 50% employee population. But I don't think those kind of statistics exist.

  • S-townMike

    Very cool, Chris. Thanks for doing this. It's really helpful.

  • Paul

    There is probably a definition of what 'crime' means in this instance and I am assuming that it falls somewhere above your normal traffic stop for a broken headlight. I do wonder if there is a tendancy in some areas for people to report crimes more than in other areas. Is there any research on that?

    The scary part of the 'per capita' issue is that eliminating people from a particular zip code (thus reducing the population) would seem to reduce the crime rate (let's assume that 'eliminating' translates into relocating a large number of people from a housing project or retirement community rather than some sort of terorist event).

  • Chris

    Well, the figure for crime I used was "UCR", which I think is an amalgamation of both violent and property crime. They also had it broken down by property vs violent crime, but I was too lazy to provide those as well.

  • Pingback: Jackson Miller

  • http://Fhl Fhl