September 9, 2013
Crime Doesn't Climb In San Francisco
Mapping the 15,000 incidents reported to the San Francisco Police Department in January and February of this year, Gordon Wintrob and Peter Reinhardt attempted to analytically answer the question: does "crime climb" in San Francisco? Or in other words, do higher elevations provide a natural barrier to crime?
Segmenting the 15,000 reported incidents based on the elevation at which they occurred, the two found that crime levels drop off sharply at higher elevations:
From the team:
One flaw in this analysis is that it could be a byproduct of the fact that there is less land mass in the city at these higher altitudes. In other words, if 90% of the city is at an elevation less than 25m, then 90% of the crime would occur at lower altitudes, assuming an even distribution of incidents. To correct for this problem, we took a distributed sample of 10,000 locations in San Francisco and divided the number of incidents in each elevation range by the number of locations in that bucket.
When normalizing for land mass at different elevations, the trend of lower crime at higher elevations was equally as dramatic:
Unfortunately the duo's analysis didn't also normalize for the difference in densities at the various elevations, a correlation which tends to be rather strong.
∙ Crime Doesn't Climb [github.com]
First Published: September 9, 2013 4:30 PM
Comments from "Plugged In" Readers
Living in the Haight-Ashbury for over 30 years, we often refer to the "Shopping Cart Line" in the same fashion the weatherman will start talking about the snow line on TV in a couple of months.
Posted by: redseca2 at September 9, 2013 5:03 PM
i guess doing a study on crime vs. elevation is less politically hazardous than, oh just a wild guess, housing project locations or demographics.
Posted by: condoshopper at September 9, 2013 5:38 PM
Correlation != causation.
Posted by: p3p at September 9, 2013 6:15 PM
But correlation may be all we care about, as I doubt that anyone is proposing raising the average elevation of San Francisco in an effort to reduce crime. On the other hand, if I am looking to live in a low-crime area, I might consider higher elevations.
Posted by: NJ at September 9, 2013 6:24 PM
I don't think they were suggesting that a low or high elevation causes or stops crime. Only that crime rates go down with elevation. The cause is probably that criminals are lazy, although there are certainly other possible causes.
Posted by: lyqwyd at September 9, 2013 6:26 PM
First, I want to give them credit for publishing their code, such transparency is rare.
Second, it looks like they didn't normalize for population density. I've contacted the authors for confirmation.
If they haven't, then this is just a waste of time, pretty as it is.
[Editor's Note: As we wrote above, "Unfortunately the analysis didn't also normalize for the difference in densities at the various elevations, a correlation which tends to be rather strong."]
Posted by: Jake at September 9, 2013 7:20 PM
There is absolutely no likelihood that the views provided by higher elevation lead to more expensive real estate lead to higher salary/education per resident lead to less crime, right?
Posted by: JWS at September 9, 2013 8:54 PM
Doesn't population density also decrease at higher elevations? How much crime occurs on Mt. Davidson for example?
This data would be more relevant if reported as "crimes per elevation level per population" instead of land area. Those results would still show a skew towards higher elevations but it might not be quite as dramatic.
Posted by: The Milkshake of Despair at September 9, 2013 10:27 PM
The Presidio gets NO RESPECT! Why I've seen what someuvyoupeople are doing out there with golf balls out on that course and I seen dogs hump firehydrants with more grace. NOW THAT'S CRIMINAL.
BIG DATA? My last wife was a big dater!
Posted by: r.dangerfield at September 9, 2013 10:38 PM
There is also very little public housing and zero SROs at higher elevations in SF. I live in Diamond Heights which might have some of the only public housing at elevation in the city, I can't think of anyplace else it exists. (And admittedly, while DH may have slightly more crime than some other hill neighborhoods, it does feel generally safe).
Posted by: Jason at September 10, 2013 10:20 AM
Projects over in Potrero are at elevation. That neighborhood is gnarly, too. Would be interesting to do this on a neighborhood-by-neighborhood breakdown. Nice that they've exposed their code. Maybe I'll play with it and see what I can tease out.
Posted by: MrsLotus at September 10, 2013 10:30 AM
"it looks like they didn't normalize for population density.
If they haven't, then this is just a waste of time, pretty as it is."
It's already been noted as an issue. The source is right there, why don't you do it yourself?
Posted by: nobody at September 10, 2013 10:33 AM
I've been saying this about Diamond Heights for years. It's nice to be vindicated.
Posted by: MysteryRealtor at September 10, 2013 2:16 PM
There are tons of projects at elevation in Bayview. Plus the aforementioned Potrero Heights projects and ones in Diamond Heights.
Posted by: NoeValleyJim at September 10, 2013 4:55 PM