We are always on the hunt for new ways to visualize the structure of the financial markets. Last week we presented an interactive correlation heatmap for the stock market. This is a great method for encoding the full correlation matrix for a group of stocks into one dataviz using color to represent the strength of the pair-wise correlations.
When our universe of assets grows beyond small bounds, we require more nuanced ways to explore the co-movements of large groups. We propose a novel approach which traverses the fully-connected correlation matrix searching for meaningful relationships:
- Select a target asset, e.g. Citigroup (C, encoded in bright red at the center of the graph)
- Search for all assets with an absolute correlation with the target asset >= a threshold value, e.g. 0.5
- Repeat the process for each asset sufficiently correlated with the target asset and graph the result:
Click to explore an interactive network graph @ mktstk.com
We can repeat this process for different target assets and correlation thresholds, allowing us to quickly identify asset clusters which are related to the target stock. Rolling over (or tapping on mobile) each node displays the relevant correlation coefficients. This should work on all platforms, but as always please comment with any bugs or reach out to us at the usual places.
Want more financial data visualizations? Sign-up to become a Beta Tester for our Corexplore project, a groundbreaking platform for financial visualization, and get free access to the platform for a limited time until the Beta Period ends. Corexplore is intended as a professional tool for Traders, Portfolio Managers, Risk Managers, Hedge Funds, and other institutions to visualize the complex correlation and volatility structure in modern financial markets.
Make sure to download our new book Intro to Social Data for Traders by our very own Thomas Pendergrass for tutorials on accessing social data sources like Estimize, StockTwits, Twitter, and Google Trends.
Categories: Quant Tools