We all didn't know each other before coming to the hackathon, so we thought to ourselves: what if we could improve the process in which people find partners/teams at hackathons? Then we realized this had a larger use case than just hackathons, and this could apply to any open source project.
What it does
You enter your github username on your phone. Our app then gives you the contact of 5 developers in your area that have the most similar development experience to you.
How we built it
We made a light-weight mobile app in React-Native for the frontend. Most of the heavy lifting is done on a node.js server running on Google Firebase, which mines Github's API for your commit history, and developers' commit history in your area. It then runs an algorithm to determine which developers are most relevant to you.
Challenges we ran into
Github has a strong limit on API requests per hour. Linkedin does not supply open access to their API
Accomplishments that we're proud of
None of us knowing each other beforehand, but agreeing on an idea and tech stack, and throwing together a full blown full stack mobile app with big data applications in less than 36 hours.
What we learned
Data mining through http requests. Front end development. Back end development.
What's next for Project Buddy
- Adding in more machine learning features outside of the difference of one's experience in different programming
- languages and creating more robust models based on labeled datasets.
- Adding a feature to match developers with designers.
- Add a feature to solely find partners/teams at a hackathon
- Incorporate devpost data into our model
- Analyze code semantics for users to match developers with similar coding styles and design principals
- Add a mentor matching system.
- Analyze Github and LinkedIn bios to gain insights into what people are 'interested' in developing, but haven't
- developed yet.
- Create our own centralized data set from sources like Github, LinkedIn, Facebook, and Devpost.