TO Community Safety Tool

• 

  Deck

• 

  Neighbourhood Incident Overview

• 

  Neighbourhood Incident Prediction

• 

  Neighbourhood IncidentForecast

What it does

Focusing on Community Safety, this project analyzes crime and socio-economic data to understand incidents for Toronto, Ontario neighbourhoods.
The data is stored in tidb serverless. The user can leverage this data to ask questions in plain text in the Overview section. It also enables the user to predict incident type and forecast incidents for particular neighourhood.
Based on the predictions and forecast it would leverage Gemini AI to generate safety tips.

How we built it

• It was built using python for data analysis, plotly for visualization, streamlit as the UI and Google Gemini was used as the LLM. Also uses tidb serverless using vector search and Google gemini to ask text based questions and get text based answers ## What we learned
• Ability to incorporate LLMs for insights ## What's next for TO Community Safety Tool
• Improve prediction and forecasting model
• Provide ability to refresh data on a weekly basis better

Built With

• gemini
• streamlit