Stock'd

Dashboard with nodes
Simulation results
AI-powered risk analysis

Inspiration

We realized that traditional investment tools measure diversification by counting stocks or sectors without capturing how risk is shared across companies. In reality, many portfolios that appear diversified are actually concentrated in a few underlying economic drivers due to hidden correlations. This makes it difficult to understand how risk can propagate across holdings during market stress. We wanted to build a platform that reveals true diversification by analyzing how risk is interconnected across an entire portfolio.

What it does

Stock’d turns the stock market into an interactive network simulation. Instead of viewing companies as isolated tickers, we model companies in the S&P 500 as interconnected nodes. Using real market data, users can introduce hypothetical events, like a major earnings drop at a certain company, and watch how the impact spreads across related companies and sectors. The platform visually demonstrates how shocks ripple through the market, helping users understand systemic relationships and experiment with “what-if” scenarios, while ensuring not to forecast the market.

How we built it

We built Stock’d using Next.js, Python, and Supabase, with financial data from yfinance. Historical stock prices were converted into returns, which we used to compute correlation matrices between companies. These correlations formed a weighted graph where each company is a node and each edge represents a market relationship. The backend processes these relationships and propagates simulated changes through the network. The frontend visualizes the network so users can see the cascading effects of market events in real time.

Challenges we ran into

One of the biggest challenges was defining meaningful relationships between companies. Raw correlations can produce an extremely dense network, so we had to experiment with filtering techniques and thresholds to highlight only the most relevant connections. Handling large datasets and maintaining responsive visualizations also required careful optimization.

Accomplishments that we're proud of

We successfully built a working system that transforms real financial data into a dynamic network of market relationships. Our simulation can take a change in one company and propagate its effects across many others in an intuitive visual interface. Most importantly, we turned complex financial modeling concepts into something interactive and easy to understand.

What we learned

Through this project we learned how interconnected financial markets truly are. We gained experience working with financial APIs, handling time-series data, computing statistical relationships between assets, and building graph-based simulations. We also learned how important data cleaning, performance optimization, and thoughtful visualization are when working with large financial datasets. There are 500 companies in the S&P 500, which means that the total number of relationships is (500) * (500) = 250,000 calculations!

What's next for Stock'd

Next, we want to expand Stock’d beyond simple correlations by incorporating sector data, market factors, and more advanced statistical models. We also plan to integrate real-time market updates, support historical event simulations, and improve the visualization so users can explore the market network at different levels, from individual companies to entire sectors. Ultimately, we want Stock’d to become a powerful tool for understanding how events propagate through financial markets.