Data Science Project Developer Independent Project
Designed and built an end-to-end earnings-call sentiment and signal analysis pipeline. Sourced and processed more than 34,000 sentences from 80 S&P 500 earnings calls, applied an AI language model for scoring, and joined the resulting data with share-price movements.
Defined metrics, event windows, and statistical tests to investigate whether management tone relates to share-price reactions. Wrote 13 automated tests, identified and corrected a data fault through raw-data validation, and excluded unreliable records.
Built and deployed a public dashboard for non-technical users. Documented a statistically inconclusive result transparently, including methodological limitations, in a README for technical and non-technical audiences.