Team Name: pApAyAS Andrew ID: allenzhe, anirudhp, soumyadc, alchen2

Inspiration

The experience of moving into college is always scary, but this is magnified when campus is in the middle of a city. This is especially true for international students, as this is an entirely new experience for them.

Google Maps can easily give the fastest route from point A to point B, but these routes could be travelling through pickpocketing hotspots or dangerous areas. This is why we came up with MUGN’T, an app that allows you to avoid high risk zones and see safer routes.

What it does

MUGN’T inputs a start and end point for a certain city, and a “safety priority coefficient” which the user decides upon. It notes the two points and runs through Dijkstra’s augmented with probabilistic algorithms to find the optimal path with respect to the “safety priority coefficient”.

The path is then evaluated with a Markov Chain that estimates the probability of a violent event occurring along the route. This is done by taking into account the historic number of violent crimes and the density of the various cities. After comparing it to the same code run without evaluating the varying rates of crime, we give the user back both the optimal path we calculate and the improvements in risk.

How we built it

The core of the project lies in the data collection and probabilistic methods used in pathfinding. We used a pair of machine learning models to extract map data and crime density data from heat maps that we found online, this was then fed into a pathfinding algorithm that prioritizes safety based on a “safety priority coefficient” that the user inputs.

Challenges we ran into

We faced a lot of challenges over the machine learning segment of our project, due to the And the front end

Accomplishments that we're proud of

We're particularly proud of efficiently extracting the roads from google maps and the success of our algorithms in finding safe routes.

What we learned

We learnt a lot about data science methods on real world datasets, as well as various tools that can be used to make a more user friendly tool.

What's next for Mugn't

We hope to integrate real time information into our system, so that we can also avoid areas with ongoing crime incidents.

Share this project:

Updates