Eric Rodrick Hu ¹ and Jonathan Thamrun ² , ¹ Fairmont Preparatory Academy, USA, ² University of California, USA

Markets move with sentiment, yet students lack a transparent, per-ticker tool to read social mood in real time. Our program, Stock Scout, ingests Reddit posts, scores them with VADER plus a small finance lexicon, aggregates daily values, and visualizes a 14-day trend in a Flutter app backed by FastAPI [2]. The three core systems are: (1) an NLP pipeline for Reddit retrieval, scoring, and daily aggregation; (2) authentication and session bootstrap with Firebase; and (3) a History Service that logs per-user searches and views in Firestore [1]. Design challenges—slang, sarcasm, rate limits, and latency—are addressed with filtering, finance-term boosts, ±0.10 thresholds, and resilient endpoints. We evaluated correctness on 20 labeled finance headlines; our classifier produced 18/20 accurate labels (90%), with errors clustered around recall and secondary-offering language. The result is a lightweight, interpretable signal that helps learners and retail investors quickly gauge sentiment and track changes without heavy infrastructure or black-box models.

Stock Sentiment, Reddit NLP, Market Trends, Financial Analytics.