Scoreboard Savant

Inspiration

After many years of sports betting on college football games I found a trend. I could never consistently win bets and make money. You know the saying "99% of bettors quit before they win big", well, I refused to be in that 99%. In order to help me and many other bettors alike, we decided to create Scoreboard Savant so that we can win big!

What it does

Scoreboard Savant makes predictions on college football games using a machine learning model. It takes data from college football games played in the last 10 years to make these predictions. It then visualizes this model using Streamlit to put the predictions in a more digestible format. Bettors can then look at these predictions in order to make more informed bets about college football games.

How we built it

We built a regression forest to make the Machine Learning model that makes predictions. We then imported Streamlit to graph these predictions and data and imported it onto a HTML site. The front end of the site is made using HTML and CSS.

Challenges we ran into

Some of the challenges we ran into were getting the correct packages to import, getting Github to work across all 4 of the developers and developing the front end as a lot of us did not have much experience with HTML and CSS. We also found trouble staying on task as we had our fair share of shenanigans in the GDC during this weekend.

Accomplishments that we're proud of

We are proud of the growth we showed during this event. For 3 of us this was our first hackathon and our first time working with some of these apps like Streamlit and react (although the final project does not use react, we tried to use it). This experience is something that we will take with us through our ventures going forward when it comes to creating similar apps.

What we learned

We learned how to use a lot of different applications like Streamlit and react. We also learned how to code more efficiently in HTML and Python. We also learned how to problem solve when our code ran into errors or was not giving the correct output.

What's next for Sports Betting

What the next step would probably be is to expand into other sports. This could be very useful for bettors that bet on NFL, NBA, or any other professional league games. Using this same regression forest model we can predict games at a fairly accurate rate and potentially win bettors more money than they would usually win.