The CrimeCategoryPrediction project uses data from the Toronto Police of Ontario, Canada, which includes information about major crime indicators (MCI) such as the date and time of the crime, and its location. The goal of this project is to predict the category of crime based on this information.
The categories of crime include Assault, Break and Enter, Theft Over, Robbery, and Auto Theft. The model uses data about the neighborhood, date and time of occurrence, premises type, and other relevant details to make its predictions.
By predicting the category of crime, one can identify patterns in criminal activity. This can help law enforcement agencies to focus their efforts on certain types of crimes or areas where particular crimes are more prevalent. Based on such information, police can increase their patrols in that area to prevent and deter criminal activity. If I can combine this model with statistics on homicide data, the usefulness should be justified even further.
This is an initial version of the model, and there is room for improvement. While working on this project, I have learned
- how to encode categorical labels
- how to handle mixed features that include both numerical and categorical data types
- how to balance your data when your have categorical labels