"Mumbai Police Embrace AI: Students Develop Crime-Predicting Tool"

Inputs: White silicon Desk

A team of final-year B.Tech students from K.J. Somaiya Institute of Technology (KJSIT) in Mumbai developed an innovative AI-powered predictive policing system in collaboration with the Mumbai Police. This tool was designed to enhance law enforcement efforts by forecasting potential crime occurrences, specifically targeting mobile theft, chain snatching, and vehicle theft, using advanced machine learning techniques and historical crime data.

Development ProcessThe project emerged from a unique collaboration between KJSIT and the Mumbai Police, focusing on improving policing efficiency in the east region of Mumbai (Zones 6 and 7), which spans 18 police stations and involves over 600 officers. The idea was sparked during an AI training session conducted by Prof. Radhika Kotecha for police officers, where discussions with Deputy Commissioner of Police (DCP) Navnath Dhavale highlighted the potential of AI in crime forecasting.

Remarkably, the system was developed and tested within just one month, reflecting the dedication and skill of the students and faculty involved. The development team comprised final-year B.Tech students Jatin Soneta, Dhruv Mokashe, Hemanshu Rajde, and Chirayu Agarwal.

How It WorksThe AI tool analyzes five years of crime records to identify patterns and predict future incidents.The Key features include:

Data Preprocessing: Historical crime data is cleaned and organized for analysis.Haversine Formula: Used to calculate spatial relationships and identify crime hotspots based on geographic data.Hierarchical Density-Based Clustering: A machine learning technique that groups similar crime occurrences to pinpoint high-risk areas.The system generates:Weekly Reports: Highlighting probable crime types, locations, and time windows.Monthly Analytics: Summarizing crime trends by area, delivered in PDF format.These insights help police optimize patrolling schedules and allocate resources effectively.Impact and AccuracyDuring its pilot phase in March 2025, the tool achieved a 75% accuracy rate, successfully predicting theft locations within a 500-meter radius and aiding in the prevention of three incidents. Additional Commissioner of Police (Addl. CP) Mahesh Patil noted its accuracy ranged between 75-80% in real-time deployment. The system specifically targets prevalent crimes like mobile theft, suspected missing mobile cases, vehicle theft, and chain snatching in Zones 6 and 7, making it a valuable asset for crime prevention and resource planning.RecognitionThe student team received certificates of appreciation from the Mumbai Police for their groundbreaking work, acknowledging their contribution to public safety and technological innovation.

Significance

This AI-powered predictive policing system showcases the potential of collaboration between educational institutions and law enforcement. It not only demonstrates the technical prowess of KJSIT’s B.Tech students but also sets a foundation for future AI-driven solutions in crime prevention. By leveraging technology to address real-world challenges, this tool marks a significant step forward in enhancing safety and security in Mumbai.