For years, the city of Chicago has a had a severe crime problem. To help combat this issue, the city of Chicago has released a full comprehensive data set containing all crime that has taken place since Jan 1st, 2001. My goal is to develop a prediction algorithm that will try and be able to tell what type of crime has been committed given certain meta data. I utilized a two different algorithms such as K-Nearest Neighbor (KNN) and a Decision Tree using the geo-location the crime occurred, and looking at a time series the crimes take place in specific months of the year to start. In the end, I was able to predict the number of each crime for a given police beat during a specific month with an r2 score of 0.54 using a KNN approach.
- Data Mining
- Machine Learning
- Geo location
- Pandas (DataFrames)
- Numpy (Math)
- Seaborn (Visualization)
- Sklearn (Algorithms)