Gameline

Inspiration

We were inspired to create Gameline because let’s face it—sports gambling is stressful, unpredictable, and sometimes downright heartbreaking. How many times have you confidently bet on a player only to see them mysteriously “resting” or have a surprise cold streak? We wanted to solve this real-world problem by giving fans and bettors smarter, data-driven insights to improve their chances of winning while still keeping the thrill of the game.

What it does

Gameline allows users to input a specific player, betting category, betting line, and an opposing team to see projected stats for that matchup. It provides detailed projections for points, rebounds, assists, steals, blocks, and other key categories, giving sports fans and bettors a data-driven edge when evaluating player performance.

How we built it

To build Gameline, we used Python and the nba_api to gather historical player and team data, including recent games, head-to-head matchups, and advanced stats. We then trained a machine learning model to predict each player’s performance in different statistical categories. The model captures patterns in player behavior and team context, enabling accurate projections for any player against any opponent.

Challenges we ran into

During the hackathon, time was a major constraint, which made it difficult to collect and process all the data we wanted and fine-tune our model. Accuracy was another challenge, as predicting player performance involves many variables, from individual stats to team matchups and game context. We also ran into technical issues connecting our backend to the frontend, which occasionally caused delays when retrieving and displaying NBA data through the app.

Accomplishments that we're proud of

We’re proud that despite the tight time constraints of the hackathon, we were able to build a working ensembled model that combines multiple machine learning methods to produce predictions that seem surprisingly accurate for player stats. Pushing through challenges like complex data variables and limited time, we successfully integrated the model into a functional app that lets users input a player and opposing team and get projected stats in key categories. Seeing the end-to-end system work under these conditions was a major achievement for our team.

What we learned

During the hackathon, we learned a lot about deciding which features have the biggest impact on machine learning accuracy, and how careful feature selection can dramatically improve predictions. We also gained valuable experience in backend-to-frontend integration, handling API calls and server connectivity to ensure the app functioned smoothly. On the teamwork side, we learned how to divide tasks effectively, coordinate our efforts, and work together efficiently under tight time constraints. Overall, the experience taught us both technical skills and practical strategies for collaborating on complex projects.

What's next for Gameline

Moving forward, we plan to focus on improving the model’s accuracy by experimenting with additional features and refining our algorithms. We also aim to deploy the project on AWS to make it accessible to a wider audience. Finally, we want to enhance the user interface to make it more intuitive and visually appealing, ensuring a smoother experience for users exploring player projections and stats.

Built With

• lgbm
• nba-api
• numpy
• pandas
• python
• scikit-learn
• typescript