Supervised ML model(classification) trained on LAPD data with attributes like crime location and time of occurance along with victim details. Random Forest and XGBoost algorithms used to determine if a given crime will remain unresolved along with the probabilty of resolution.
- Decreasing trend in resolved crimes (green) and icrease in unresolved crimes(red)
- Ratio decreases from 1:3in June 2018to 1:7 in June 2023
Location based crime analysis (link)
Tableau was used to visualize year-on-year resolution percentage across different counties of LA. Refer to link for intercative graph.
Analysis based on Victim details (link)
The above visualizations clearly showed the growing disparity in resolved and unresolved crimes on the basis of different attributes like location, premise, weapon use, race and age.
