Like many people, I get the majority of my daily news from online sources. A few days ago I noticed an interesting computer-related news story, and clicked on the link to the story. The story involved a computer event in the city of Chicago, and originated from the Chicago Tribune newspaper.

As I finished reading the short article, and scrolled to the bottom of the page, I noticed that the Tribune had posted their weekly homicide map update. What a minute. What? A weekly homicide map? Yes.

Chicago Weekly Homicide Map

The notation next to the map boasted that in 2017 only 636 people had been murdered through the first week of December. I found it unusual that only 636 murders in a city is something to celebrate. I dug a bit deeper. According to a few follow-up Internet searches, the most dangerous street in Chicago is the 6400 block of South Martin Luther King Jr Drive. However, entire sections of the city have a huge number of homicides and other crime.

I did a few other searches for crime maps. The nearest large city to me is Seattle. It has relatively little serious crime, but a lot of car thefts and burglaries, especially near the University of Washington. Makes sense I guess.

During my searches I found homicide and crime maps for a few other U.S. cities. Cities that seemed especially bad were Baltimore, Detroit, St. Louis, Oakland, Atlanta, and Memphis. I haven’t been to any of these cities and don’t know them, so I don’t know why they have so much crime. Although I couldn’t get concrete data, an Internet search for the most dangerous street in the U.S. returned Martin Luther King Jr Blvd in St. Louis as a strong contender.

St. Louis Homicide Map

Baltimore Homicide Map

Anyway, the point is that geolocation data can be a powerful tool. Currently, crime maps just identify the locations of past crimes. Machine learning is all about prediction. Perhaps machine learning can be used with geolocation data to predict and reduce crime. No city newspaper should have to publish a weekly homicide map.

I’ve worked with several geolocation code libraries, including Google Maps, Bing Maps, and Leaflet.js and in my opinion, Bing Maps is by far the best. See https://msdn.microsoft.com/en-us/library/dd877180.aspx and my free e-book “Bing Maps V8 Succinctly” at https://www.syncfusion.com/resources/techportal/details/ebooks/Bing_Maps_V8_Succinctly.