The Crime Landscape of St. Paul, Minnesota
Anonymous Student
The purpose of this project is to analyze crime-related data from St. Paul, Minnesota. Focus is not only given to analyzing patterns seen in St. Paul's crime, accident, and traffic history more broadly and overtime, but also spatially across St. Paul's Police Grid. Other information, such as liquor licenses, vacant buildings, and bike paths, is used to predict the number of crimes in a given area. These analyses and findings are used to make recommendations to the St. Paul Police Department on possible actions they could take to better respond to the crime climate in St. Paul, as well as priorities and behaviors that need to be reevaluated.
The results of this project include the following:
- The total number and distribution of crimes in St. Paul have not significantly changed over the past five years, with Assault, Domestic Violence, and Discharge incidences making up over 60% of crimes each year.
- The spatial breakdown of crimes across the St. Paul police grid has been stable over the past few years, with a large number of crimes occurring along University Avenue, in downtown St. Paul, and in the Payne-Phallen neighborhood. Crimes are most likely to occur during the summer months and December.
- For the past ten years, approximately 80% of annual traffic stops have involved White or Black drivers, with Asian, Latino, and Native American drivers making up the rest. The percentages of traffic stops involving White and Native American drivers have been relatively stable, while Black and Latino percentages have been falling. The percentage of traffic stops involving Asian drivers has been steadily increasing by 0.25% of total annual traffic stops per year. Black drivers are the most over-represented ethnicity with regards to traffic stops, representing nearly 40% of traffic stops while only making up 16% of the population. While population data helps to explain some of these trends, others appear to be the result of biases in the way the St. Paul police department patrol the city's drivers.
- Each ethnicity's traffic stops are focused in anywhere from one to four areas around St. Paul, with each ethnicity's most active ten precincts out of two hundred making up anywhere between 20% and 30% of annual traffic stops for the given ethnicity.
- The number of liquor licenses in a given precinct has little to no linear relationship with the amount of graffiti in the same precinct, as the two variables have a correlation of only 0.05128.
- Many of a given precinct's characteristics are moderately positively correlated with the annual number of crimes in the same precinct. The number of vacant buildings and number of crashes in a precinct, for example, have correlations of 0.408 and 0.736 with the number of annual crimes in the same precinct.
- Despite many of the variables have moderately strong correlations with annual crime rates, a Least Squares regression model including annual number of crashes, traffic stops, vacant buildings, bikeways, and liquor licenses does a poor job of predicting the number of annual crimes in a given precinct. Steps to improve this model include collecting more data from different years, including additional variables not currently tracked by the St. Paul police department, and exploring more complicated model types.
To learn more about the data used, click here.
To read about acknowledgements, click here.
To view the analyses and results, click here.