AI Event Matcher

Inspiration

We all came into Santa Cruz with no experience about any clubs and honestly Cornucopia was a bit overwhelming for everyone. There are hundreds of clubs and there are so many people so finding and talking to every club is impossible. We thought why not automate it and put it on a website for everyone to see.

What it does

Our application has 2 aspects to it. One is a discord bot that scrapes data from the server and uploades it into a CSV File in which it has all the information about meeting times, events, etc. This is the data that the Chatbot that we created uses to answer all of its questions. We also have a Web UI part for users to see different clubs that are for them. They can go on here choose some tags and then continue looking at the clubs and upcoming meetings for these clubs.

How we built it

For the front end, we had lovable build it for us since we didn't know any front end. For the back end, we used a ChatOllama model and then created an agent from the data we scraped through discord.py.

All the code that we wrote was python code - discord scraping, chat bot, etc.

Challenges we ran into

We had to learn a lot of new API and documentation - discord.py, Chatbot (how to create an agent), etc. This took a bunch of our time since this is our first time working with these tools.

We also didn't really know how to make a very good front end. We know a little bit of HTML/CSS but nothing enough to create a good website so we ended up vibe coding the whole front end.

Another problem is that not every server uses discord so it would be hard to make this something universal.

Accomplishments that we're proud of

We are proud of what we created so far. We came into this not knowing anything about how to create anything like this and we are just happy that we were able to scrape data on a mass scale through the bot we created and also being able to use that data as a training model for the chatbot that we also made. We do wish we had a bit more time in order for us to maybe look into image recognition as well but we're proud of the product we made.

What we learned

We all definitely learned that this is a team event, and not everything can be done by one person. One important thing that we have to think about is how to schedule our time wisely. We also have to learn how to split up the work equally amongst one another.

This hackathon also gave us a decent exposure to pushing a product out on time and really tested how fast we could build things.

What's next for AI Event Matcher

As I mentioned above, our plan is to continue building this and go from text detection to also detect flyers and different images. We want to work with scraping off instagram pages and make it so that it actually displays on our front end. To do that though we need to learn a lot of front end technology so that will be where we will start first.

Built With

• css
• csv
• discord.py
• python
• typescript