Mohammed A. and Yael F. repository for collaborating on Midterm-Project
Implement targeted, data-driven measures to address different crime types when they are likely to occur in Toronto.
Predict crime type with a classification model and find features of significance that correlate with various crime types.
- Download the "Major Crime Indicators" dataset from the Toronto Police Services website and join it with the "Toronto Neighbourhood Demographics' Cean Data" dataset.
- EDA analysis, clean the dataset and try to find patterns in data between crime type/count and features by using a statistical model.
- Build the model, create multiple classification models and compare them to determine the best.
- Visualization, create a dashboard using Tableau to present our findings and build the story.
- Random Forest has the best result with 61.9% accuracy.
- Decision Tree comes next with 60% accuracy.
- KNN was last with only 59% accuracy.
The dataset was too large, and we couldn't apply API to it due to free usage limitation
Join the dataset with other open Toronto datasets and try to improve our model.