End-to-end data science project that incorporates various elements inlcuding raw data processing, NLP-related tasks such as sentiment analysis, classification and topic modelling. Will eventually be extended to deployment via FastAPI as well as GPT-based content generation.
Objective:
- Scraping of forum posts from various subforums on https://forums.hfboards.com/.
- Analysis of the sentiment of different fan groups based on team-specific subforums and the development of their respective sentiments over time.
- Generation of follow-up contributions using GPT.
- Identification of trends in which teams are most discussed, their related sentiment, and prediction of how sentiment shifts before and after games or major events (trades, playoffs, etc.).
- Deployment of a classification model using a cloud service.