Paw Patrol

What it does

Telegram Bot - Uses machine learning to determine areas with high crime risks and provides patrol officers with the crime hotspot areas within their jurisdiction as well as the order in which they should pass by these areas to minimise travelling distance. This allows patrol officers to prioritise high risk areas, increasing the likelihood of them patrolling near potential crime, in turn reducing emergency response time. The telegram bot is a user friendly platform for patrol officers to generate the route just before their patrol duty.

Web Application - Tracks the number of active patrols for each police center and provides a map of current crime hotspots based on live incident reporting to help Police HQ allocate manpower where necessary. If there are a lot of incident reportings within an area, the map would highlight the area as a high crime risk area for the police force to focus on. This ensures efficient resource allocation based on live incident reportings

How we built it

backend

Data Collection: Did webscraping using the BeautifulSoup library to scrape the publicly available data off websites and geolocation using the GeoPy Nominatim geolocation service to retrieve coordinates of locations, which are then processed and combined using Pandas

Data Analysis: Used Random Forest Regression on part of the dataset to determine the correlation between the various parameters and crime rates in Singapore, in which the data is then extrapolated to determine potential crime hotspots around Singapore

Used K-Means Clustering with each NPC acting as Weighted Centroids to determine the jurisdiction of each NPC over the determined crime hotspots

Routing: The idea of Travelling Salesman Problem while minimising the distance travelled can be applied when determining the optimal route a patrol car should take to visit all crime hotspots under their jurisdiction when leaving from a specified NPC

frontend

The Web App used React.js as the main framework and Express.js for some of the backend functionalities such as fetching data from mySQL databases.

python-telegram-bot library to develop a telegram bot

Challenges we ran into

Unable to find recent/live crime rate data, which might be unavailable to the public.
time restraints

Accomplishments that we're proud of

As this is our first hackathon, we're proud of making a finished product even though we were unable to include all the features we initially intended to have due to time constraint.