Market Analysis System (MAS)

Inspiration

Wanted an easy way for non market investors to ask “what’s up with this stock?” and get price, chart, and sentiment in one place.

What it does

Chat-style UI that looks up a ticker, returns live quote, recent chart, and headline sentiment, then summarizes a takeaway. Additionally, we have a recommendation agent that helps the user with personalized recommendation on markets.

How we built it

Frontend: React/Vue chatbot; backend: SAM WebUI gateway with a custom stock agent (yfinance + matplotlib for chart, headline heuristic for sentiment), wired via JSON-RPC + SSE (Server-Sent Events).

Challenges we ran into

Handling proxy for the gateway, plotting charts headless with matplotlib, avoiding rate limits, and making sentiment useful with lightweight signals.

Accomplishments that we're proud of

Seamless Data Integration: Successfully engineered a real-time pipeline that fetches ticker data, generates custom visualizations, and performs sentiment analysis in a single, fluid user request.

Intuitive "Zero-Barrier" UI: Developed a chat interface that translates complex financial data into digestible takeaways, making market insights accessible to non-investors.

Robust Backend Architecture: Implementing a stable JSON-RPC + SSE (Server-Sent Events) setup allows for responsive, "streaming" updates that make the chatbot feel alive and interactive.

Automated Visual Insights: Mastering headless matplotlib generation to ensure users receive high-quality, up-to-date charts without manual oversight or server-side lag.

What we learned

The Nuance of Sentiment: We discovered that raw headlines aren't enough; weight and context matter. We learned how to build heuristic models that filter "noise" to provide a more accurate market "vibe."

Asynchronous Communication: Gained deep experience with SSE and streaming protocols to handle the latency inherent in financial API calls, ensuring the UI remains responsive while data fetches.

Complexity with Clarity: We learned the "less is more" principle in FinTech, how to take a mountain of yfinance data and distill it into a three-sentence takeaway that a beginner can actually use.