DriverDirect | Devpost

Home Screen
Emergency
Heat Map

Inspiration

With a growing amount of accidents on the road every year leaving many families in despair, we intend to mitigate this enormous problem by spreading awareness and incentivizing people to drive safer. In addition, we resonated with State Farm's mission statement of "Living Life Confidently" and intend to make sure that people will feel confident whenever they drive.

What it does

The app, using Google Maps API, determines when a person is in a accident heavy zone and will send a notification alerting the user to drive safe and be cautious. Additionally, in the very small case that the user ends up in an accident, the app will automatically call 911 and message State Farm about the situation.

How we built it

For the front end, we used Flutter with Android Studio in order to develop an Android and iOS application simultaneously. In order to develop the machine learning model, we used Python to select the columns needed from the data from StateFarm and created a heat map through GMaps. Java was used to develop the algorithm of determining the region that was unsafe.

Challenges we ran into

The main challenge we faced was figuring how to use the widgets in Flutter to build the UI for the app and how to integrate the Google Maps API into the body of the app. In addition, actually implementing the machine learning model and the algorithm to determine when to send a notification and coordinating with Flutter's widgets was extremely difficult.

Accomplishments that we are proud of

With the help of mentors, we are proud of developing a machine-learning model in order to determine where the accidents are most concentrated and an algorithm to determine how to fix each separate point into a uniform field where people would be notified to drive slow.

What we learned

We learned how to use Flutter's containers to build a body which used Google Map's API. In addition, we learned how to better scrape the data in order to build the machine learning model and the Java algorithm.

What's next for DriverDirect

We would like to integrate our machine learning model with State Farm's Drive Safe and Save App so that we could help people realize not only to drive better, but to simply either avoid or be extremely cautious of area with a high amount of accidents based off numerous variables from the time to the weather at the time of the crash. In addition, we would like to integrate our heat map model and warning notification system into Tesla's onboard UI system.

Built With

Submitted to

HackSMU II
- Winner Best Pitch - Amazon Echo

Created by

I worked on the data analysis aspect of the project. I mainly used the google maps API, Pandas, and Numpy to generate a heat map of the data so that an algorithm can be developed. In addition, I cleaned the dataset so that it can be fed accurately into the machine learning model. Simultaneously, I used the sci kit-learn machine learning library to make clustering predictions on the data set. Finally, I also used the machine learning model and converted it to a readable file for the application.

Pranay Lakkaraju
I worked on developing the algorithm to determine where an area is accident prone and also determined the parts of the data that was important for developing the machine learning algorithm. In addition, I worked on developing the home page for the app through Flutter and also integrated the machine-learning model and algorithm into the app.

Dinesh Polisetty
Abhinav Kolli

Updates

Dinesh Polisetty started this project — Feb 02, 2020 10:18 AM EST

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.