Inspiration

Daniel is a huge fan of VCT, going as far as to staying up from 4am to 8am to watch the Swiss stages play out. While discussing with the group of potential ideas, in some random sequence of events, Min brought up basketball. Some dots connected in Daniel's brain and he thought of fantasy Valorant. Similar to how fantasy basketball works, but, obviously, with Valorant Players.

What it does

Just like any other fantasy sports, you create a team of five players. These players could be your favourite players, a dream team, or a team full of KangKangs. However, if you are on the more inexperienced side, our chatbot, named BoomBot, will help you compare players, summarize a player's strong points, to tell you who the best Jett main is.

After choosing 5 players, you have a chance at having your team compete in a replica of VCT Champions 24 and maybe try to take the trophy for yourself!

How we built it

After making a Figma design for our frontend developer, who used React.js, we began working on scraping data using Python and BeautifulSoup4. After scraping all the data we needed, we uploaded all the necessary information into MongoDB. We then began working on API endpoints using Flask and building our chatbot and our AI to determine the winner using Cohere and AWS bedrock. As far as making our chatbot ACTUALLY smart goes, we setup a RAG pipeline for the bot so that it can take specific necessary data into the context when answering questions, rather than the entire context, which can be FAR slower and FAR more costly!

Challenges we ran into

The main challenge was trying to finish this project on time, while dealing with University school work. Both Fahmi and Ruby have to write a midterm (which historically averages 50% or some reason) the same day this is due, so balancing this hackathon and school is definitely the biggest challenge.

Accomplishments that we're proud of

Finishing this ambitious project, though we did have to leave a couple of features at the door.

What we learned

Never procrastinate hackathons to do AFTER reading week.

What's next for Vantasy

The biggest issue with Vantasy would be how winners are determined. The hard part about a project like this is accounting for non-quantitative statistics. For example, someone like EDG nobody, who is the IGL for EDward Gaming, led his team to winning VCT Champions 24. Not to mention, that was his first year being an IGL after replacing the former IGL. Despite these amazing feats, his stats are not as good compared to his peers, which the AI might look at it as "nobody is not a great player." So the first thing Vantasy needs to improve on his somehow incorporating more than just numbers into calculations.

Built With

Share this project:

Updates