We hear stories about crimes happening around us everyday. It's especially terrifying since we're students and we just want to enjoy our college experience. While discussing ideas to hack, we realized this problem resonated with all of us. We immediately realized that we need to work on this.

What it does

It gives you and your loved ones a chance to take your safety into your own hands. By using our application, you are taking control of the most important thing you have, your life. It generates the safest and shortest path from Location A to Location B.

How we built it

We analyzed large data sets of crime over the past 10 years, picked up representative points after weighing them based on relevance and severity of crime. We then created a data anaytics flow (that can be easily repeated with new data) to glean more useful insights from the raw data. To help analyze the data further we plotted the data on Virtual Reality to look at data spreads and densities. This processed data was than pushed into our MongoDB database hosted on MLab. We then used these generalized clusters to see which areas we should avoid while routing using the Wrld and OSRM API's. From here our app generates a single route off of all the information based on what our algorithm devises to be the "safest route" for the user. This route is then plotted and displayed for the users convenience.

Challenges we ran into

We started out with millions of cells of data, which made it really hard to work with. We had to filter out the most relevant data first and convert the data into a completely new format to optimize it for our path finding algorithm. In addition, we had to optimize our MongoDB to work well with our data as we made frequent queries to a database with originally over a million elements. In, addition, we constantly had issues mixing up latitude and longitude values as well as making HTTP requests for many APIs.

Accomplishments that we're proud of

We are happy that we were able to create a minimum viable product in this short duration. We're especially glad it's not just a "weekend-only" idea and it is something we can continue to make better. This idea is something we hope in the future can be something that truly has a social impact and can actually make the world a safer place.

What we learned

  • Wrld/OSRM API
  • Node/Express/MongoDB -HTTP Requests/Callback functions -Unity VR sdk -Pandas/Numpy
  • Coordinating team efforts in different parts of the Application

What's next for SafeWorld

Using real time data to better predictions for real life incidents (protests). In addition, having more global data sets would be a optimal next step to get it working in more cities. This would be expedited by our adaptable framework for data generation and pathing.

Built With

Share this project: