Matchathon

Forming and discovering teams at hackathons can be messy and cumbersome. Therefore, we have developed a tool that helps participants cut through the noise to find team members and teams more easily.

Comment

Inspiration

Our inspiration for this tool was our own struggle at HackZurich with forming a team. The information is spread over DevPost, Slack, and the people meeting physically, which makes it hard to find teams or team members. In addition to this, the interests of a participant may change due to workshops from the challenges.

What it does

We are helping HackZurich participants signal their needs and haves with a web-based platform and make recommendations about conformity between participants and teams. Our real-time form, a participant or team can filter through the noise in different information channels. Based on this information that is used as a filter, the software is making real-time recommendations on which team or participant to contact.

How we built it

The web-based application analyzes unstructured Slack data with NLP and adds it to the structured data where applicable. Between the teams' needs/haves and the members' needs/haves, the matching engine is using a simple heuristic to identify the strength of the matches. This information is then passed on to the web front-end and visualized in three buckets of potential matches: 1) top matches (counterparts with high conformity). 2) other matches (counterparts with some conformity), and 3) all (all other participants/teams)

Front-End Built in react, with UX in mind, allowing for minimum effort to get matches

*Back-End Built in nodejs connected to a mysql database

NLP Planned for future release to include NPL and parse out data from slack messages into the current database, giving even more accurate matches

Matching Engine The matching engine is based on a point system, where matches are ranked up based on how many categories it find a match for

Challenges we ran into

Besides regular connectivity issues. The main challenge was to get a good algorithm that would produce useable results. Relevant results

Accomplishments that we're proud of

Getting the infrastructure to play well, with our deployments

What we learned

To stay hydrated

What's next for Matchathon

The software could be integrated as a new feature into DevPost to make matchmaking at hackathons easier or further developed into a standalone product.

Share this project:

Updates