A few years ago I working on a project that required plotting data onto a map. I got quite good at programming the Bing Maps and Google Maps APIs. In fact, I wrote a book about Bing Maps that I’m quite proud of: https://www.syncfusion.com/ebooks/bing_maps_v8_succinctly

On a recent weekend, I came across a news article that showed crime data placed onto a map. I thought I’d see what current crime-mapping systems are available. I found two interesting Web sites.

The crimereports.com site lets you enter a geographical location and then it plots icons for crime. Here’s the disaster of a city named Chicago showing 830 major crimes in a two-week period in what is considered the “nice” part of the city:

Chicago: A crime-infested war zone? Or is this misleading data? Answer: the map shows only reported crimes so the actual number of crimes is far worse. Over the past year, even the “nice” parts of Chicago have become extremely dangerous.

Another site is crimemapping.com and it’s similar. Here’s the pit called San Francisco, showing 803 serious crimes in a small but representative part of the city for just one week! A red pin with a number indicates multiple crimes at that location. Incredible.

San Francisco: The amount of crime on a daily basis is hard to believe.

I find mapping data interesting but, in the end, unsatisfying. Data is just data. These maps don’t tell you much except that there’s a huge amount of crime in Chicago and San Francisco. By the way, to be fair, I crime-mapped a large city near where I live (Bellevue, WA) and zoomed to the same scale/resolution/area, and there were only two crimes in a two-week period: one shoplifting and one car break-in.

Let me try to put things another way. I am fascinated by all forms of applied mathematics. But ordinary traditional statistics requires human interpretation. I’m much more interested in machine learning systems that make predictions. For example, I’d like to see a map that predicts crime in an area for next week, not a map that just shows crime from last week.

Update: The day after I finished writing this blog post, a news article appeared detailing how Shane Colombo, age 25, was murdered in the exact area shown on my Chicago map. Colombo had arrived in Chicago just four hours before his death. He came from New York and was getting ready to start graduate school at Northwestern University. He was on an errand to buy clothes hangars. According to the report, he was killed for no apparent reason at 8:25 PM by two unidentified dark skinned men, who were seen on security video running away. Could this horrible incident have been prevented if Colombo had seen a Chicago crime map and realized the danger? Nobody will ever know, but this is a sad story. One of my dreams is that machine learning and AI can help prevent things like this murder.

Graduate student Shane Colombo. One of the shooters.