One of our team members broke his collarbone falling off his bicycle. After he told us about the incident, we found the North Carolina open data for all recorded bicycle crashes. This included things like alcohol use and speed at which the car was going. We used many of these parameters to feed our custom machine learning algorithm to find out which route is safest.
What it does
This app allows bikers in North Carolina to see where the not so safe areas to bike are and avoid them with an interactive maps app.
How we built it
Using React-native,Node.js and tensor flow we predicted the most likely places for bicycle crashes to occur using the North Carolina bicycle crash data set. We implemented google maps api to feed us routes and pick the safest to navigate through our app.
Challenges we ran into
- Determining if accidents we on a given route
- Giving weight to accident characteristics to determine the severity
- Looking through the Google Maps documentation was time consuming
- Downloading Android Studio and getting it up and running took a while
Accomplishments that we're proud of
- The user gets multiple bike routes from point A to point B and is shown where accidents have happened
- We used machine-learning to determine the weight of accident characteristics
- Its a scale-able app if we had more datasets from other locations
What we learned
We learned how to use Android Studio as well as the React-Native framework. We also learned how to create a machine learning algorithm to find the safest routes from routes that google maps provided.
What's next for Cyke
Ideally we want to use data from other states so the app isn't only able to be used in North Carolina. We would also like to implement the open data for bike racks to find the closest paring spots after someone reaches their destination.