Inspiration
As a retail trader frustrated by missing market shifts, I saw a gap: social sentiment drives price moves, but no simple tool existed to track it in real time. Watching Reddit and X threads move stocks while traditional tools lagged sparked the idea to build a focused, trader-first sentiment platform.
What it does
StockSentiment AI scans social platforms (Reddit, X) and financial news to calculate a 0-100 Fear & Greed Index for US/EU stocks. It visualizes sentiment vs. intraday price trends, alerts to abnormal spikes, and lets traders backtest historical sentiment patterns against price performance.
How we built it
We started with core sentiment logic: aggregating source data, weighting positive/negative mentions, and building the 0-100 scoring system. Then we built a React frontend for the dashboard/charting, Node.js backend for data processing, and integrated Chart.js for dual-line price/sentiment visualizations. We prioritized responsive design for desktop/tablet use, aligning with Western trading platform UX norms.
Challenges we ran into
Calibrating the sentiment score to avoid overreacting to noise was tough—balancing speed (real-time updates) with accuracy took weeks of iteration. Also, building dual-axis charts that scaled correctly for both sentiment (0-100) and currency prices required custom Chart.js configurations to keep data readable.
Accomplishments that we're proud of
We created a intuitive Fear & Greed gauge with color-coded zones that traders instantly understand, and a seamless flow between the dashboard watchlist and individual stock pages. The real-time source breakdown (Reddit/X/News) gives traders transparency into what’s driving sentiment shifts—something rare in off-the-shelf tools.
What we learned
We deepened our understanding of behavioral finance and how retail sentiment correlates with short-term price action. Technically, we mastered real-time data state management in React and learned to balance data freshness with frontend performance for smooth user experience.
What's next for this project
We plan to expand the source pool to include regional EU news outlets, add granular alert customization, and integrate portfolio-level sentiment tracking. Long-term, we’ll refine the scoring algorithm with more historical data to improve predictive accuracy for price-sentiment correlations.
Log in or sign up for Devpost to join the conversation.