Market Confidence Analyzer w/ Polymarket Sentiment Analysis

Inspiration

We wanted to try something completely new that neither of us would have prior background knowledge around. Kellen pitched the idea of a market confidence tracker, and we quickly realized that it would fit perfectly for the data science track, and more importantly real-time Polymarket data would be perfect to set it apart from anything else currently available in the realm of market monitoring.

What it does

Our code utilizes data from Yahoo Finance (yfinance) for stocks, sectors, housing, bonds, and USD, CoinGecko for cryptocurrency, and Polymarket’s Gamma API for prediction markets as well as our custom formulas to create a market confidence score. The market confidence weighs the safety of investing in the stock market, and using this data, we calculate a real-time 0–100 market confidence score. Additionally the page holds static data from the past ten years for quick comparison to past trends.

How we built it

We built a complete backend and frontend system. We used Python and fastAPI to build a fully modular src/ directory in the following format: config/ → asset definitions, categories, risk-on/off direction, Polymarket market configs services/ → data fetching, indicators, Polymarket parsing, caching, confidence engine routes/ → API endpoints (/confidence, /assets, /polymarket) utils/ → HTTP helpers, normalization logic index.py → FastAPI initialization + CORS

After issues with our first code, we built the new frontend in a single html/js file that calls our fastAPI endpoint, displays the full UI, and includes a community confidence rating where users can show how well the confidence meter matched the results of their trades that day. We used Uvicorn for live backend reloading during development, and caching to avoid rate-limiting APIs.

Challenges we ran into

About five and a half hours into the hackathon after making great progress, we could not do anything to resolve a problem with the code; we made the difficult decision to start over from scratch, but thankfully had a good idea of what needed to be done, and we built the new frontend code to a single html/js file for easy validation and problem resolution. Around 6am this morning, one of our computers github stopped cooperating despite multiple experts attempts to resolve the issue, so for the remainder of the hack we were reduced to one computer for code. The biggest challenge was figuring out how to use and integrate the APIs, as neither of us had any experience in that realm.

Accomplishments that we're proud of

We are super proud of the accuracy and real application ability of our data; being that neither of us had prior experience or knowledge around the stock market, the results and capabilities of our app exceed what was planned and predicted. We are also especially proud of our resilience dealing with github difficulties and and a complete restart 5+ hours into the hack.

What we learned

Kellen greatly improved his limited frontend and html knowledge and Lake learned python for the first time. We both vastly improved our knowledge and abilities in github, and it was our first time using and integrating APIs into a project.

What's next for Market Confidence Analyzer w/ Polymarket Sentiment Analysis

We would like to add volatility control for heightened affect on confidence during highly volatile markets. We would like to create another page that focuses specifically on crypto with its own confidence and sentiment. We would also like to talk with someone who has more market experience and knowledge in order for us to fine tune what the users might want added and/or altered, as well as fine tune weights and parameters for each category.

Built With

• coingeckoapi
• css
• fastapi
• gpt5.1
• html
• javascript
• polymarketgammaapi
• pydantic
• python
• uvicorn
• vscode
• yfinanceapi