Stage 1 Report (Weeks 2-4):
At this stage, our team worked with our client to define key problems and the initial solutions for this project.
- Problem Identified: Inexperienced sports bettors are relying on intuition and bias
- Proposed Solution: An AI-Powered platform that combines both statistical and sentiment analysis
- Tech Stack:
- Frontend: Angular
- Backend: FastAPI
- Database: PostgreSQL
- Looking for a free API that can get us team, game, and live stats
- Features to implement:
- Display NBA game data
- Build prediction models
- Develop sentiment analysis system
- Develop an automated data pipeline
Stage 2 Report (Weeks 5-9):
At this stage we mostly focused on our system's design and architecture.
- Created system architecture diagram
- Designed data flow between our frontend, backend, and database.
- Built our screen flow diagram with Excalidraw
- Defined key API endpoints for our backend
- Defined database schema (games, teams, users, etc)
Progress:
- Full stack application has been initialized and deployed to server
- Built fully functioning user authentication system
- Designed frontend templates/components to easily integrate our data as we build out our API
Challenge(s):
- We do not have a reliable API at the moment.
Stage 3 Report (Weeks 10-12):
At this stage we have made a lot of progress in implementing features.
Completed:
- Found and integrated a reliable, free API for NBA data (NBA_API)
- Built out backend endpoints for game information
- Automated our data pipeline with Celery to send live data periodically (to implement live game scores)
- Developed two machine learning models for pregame and live game predictions
- Developed our web scraping and sentiment analysis systems
In Progress:
- Feature Engineering for better pre game accuracy
- Incorporating our sentiment analysis scores with our pre game model
- Just started designing bet tracking system
Challenges:
- Figuring out how to navigate the NBA_API
- Developing the sentiment analysis system with Gemini (key issues)
- Bugs appearing with displaying proper information
Stage 4 Report Weeks (13-15):
At this point we needed to build the bet tracking system and incorporate the sentiment analysis with pre game predictions.
Completed:
- Prediction model are using sentiment analysis scores
- Fully functioning bet tracking system
- Polished UI (lots of time focusing on bug fixes)
For most weeks our team meets online to discuss progress and tasks to handle, updates are given through our group chats. Weeks leading up to a demo we meet in person at the library to finish any work that needs to be done, set up our presentations, and practice them. Our team also meet every other week with our TA either in person or online, and we met with our client periodically to give updates and receive feedback.
Timeline: