100 - Jamaica Criminal Intelligence Systems

JCIS: Where criminal intelligence meets impact— using advanced machine learning to turn data into deployment, insight into action, and making every officer count in Jamaica's fight against crime.

Comment

Inspiration

Jamaica's economic potential has been held back for a long time due to crime destabilizing our social and economic environment. If we can reduce crime it will open the door to a lot of potential productive opportunities which as of now lay dormant. Our government wants to reduce crime but are hampered by the constraint of limited resources. Therefore if we can find a way to empower law enforcement to fight crime more efficiently without increasing the resources deployed it will potentially be a massive benefit for our country.

What it does

The Jamaica Criminal Intelligence System is designed to collect criminal intelligence reports from law enforcement across the island. It aggregates the data in a central location and then predicts the potential for crime in a given area through the use of a heat map. It then uses AI to give optimal recommendations to the police in terms of how they should reallocate their man power. The system gets more powerful the more data it gets and is automatically both updating its outputs and the model its using to predict crime.

How we built it

We developed the Jamaica Crime Intelligence System using the tech stack with Next.js, React, Tailwind CSS, Fast Api and Postgresql; creating a responsive and secure platform optimized for law enforcement needs. The architecture centers around a Random Forest Classifier and we utilized a component-driven approach to bring the UI to life. Our frontend was built with shadcn/ui components, custom SVG maps for geographic visualization, and Recharts for data analytics. We implemented role-based access control through a custom authentication system, ensuring officers can only access information appropriate for their rank and jurisdiction.

Challenges we ran into and what we learned

A significant challenge faced was that we had to learn on the fly how to manage trade offs between iterating, testing and training models with limited compute. Another problem we had to navigate was how best to set up out Deployment

Accomplishments that we're proud of

We managed to Create a full stack real time AI application which was deployed

What's next for Jamaica Criminal intelligence System?

Built With

Share this project:

Updates