A couple of us use Tinder, but are always scared of meeting people who are nothing like what their profiles say. So we decided to integrate live video chat so that potential matches can screen each other before going out! We match people with machine learning.

What it does

CoffeeDate uses machine learning to match you with your ideal partner. If both of you match, we enable you to enter a live video chat with each other where you can chat about a date.

How I built it

We used Angular2 for the frontend, a Flask server on the backend. We used boosted random forests utilizing a combination of the Jaccard coefficients in dating profile matches and a binary classifier to output predictions on who users would match with, which was 83% accurate

Challenges I ran into

Accomplishments that I'm proud of

Creating and integrating the dataset for the ML aspect

What I learned

What's next for CoffeeDate

Built With

Share this project: