LLM
NLP
Streamlit
OverPrompt is an LLM-powered analytics assistant built on more than one million ball-by-ball records. It accepts natural-language questions, identifies the required analysis and returns answers grounded in the underlying data.
- The LLM interprets the question and selects the correct analysis tool
- Python and Pandas calculate the numerical results
- Phonetic matching, fuzzy matching and embeddings resolve player and team names
- The application supports comparisons, trends, rankings and match-level analysis
- The interface is deployed as an interactive Streamlit application
The language model does not invent or calculate the statistics. It interprets the request and chooses the appropriate tool, while the numerical answer is calculated directly from the source data.
Tech stack: Python, Pandas, PyArrow/Parquet, rapidfuzz, jellyfish, sentence-transformers, OpenAI/Gemini, Streamlit.