Asset Correlation Matrix

📊 What is this dashboard?

The Asset Correlation Matrix shows how closely pairs of ETFs move together over a selected time window. It covers 70 ETFs spanning US equities, sectors, international markets, bonds, commodities, currencies, and thematic strategies.

Correlation ranges from -1.0 (perfect inverse relationship) to +1.0 (perfectly in lockstep). A value near 0 means the two assets move independently of each other.

🎨 Reading the color scale

Dark red cells indicate assets that move closely together. Dark blue cells indicate assets that move in opposite directions. White/light cells indicate little relationship.

📅 Time windows

Use the toggle to switch between two lookback windows:

• 1 Month (21 trading days) — medium-term correlations, useful for portfolio construction and risk management. Most stable and reliable.
• 1 Week (5 trading days) — very short-term correlations, highly sensitive to recent market events or news. Can change rapidly.

Start with the 1-Month view for strategic decisions. Use 1-Week to check if short-term behaviour has diverged from the norm.

📈 Rolling correlation chart

Click any cell in the heatmap to open a lightbox showing the rolling 21-day correlation between those two ETFs over the past 91 trading days (~4.5 months). This lets you see:

• Whether the current correlation is typical or an outlier
• How stable the relationship is over time
• When correlations spiked (e.g., during a market selloff, correlations often rise toward 1.0)

💰 Practical uses

• Diversification: Look for pairs with low or negative correlation to reduce portfolio volatility.
• Concentration risk: High correlation (+0.8 or above) between two holdings means they move together — you may not be as diversified as you think.
• Hedging: Assets with strong negative correlation can act as partial hedges. For example, GLD (gold) and TLT (long bonds) often negatively correlate with equity ETFs in risk-off environments.
• Regime detection: When correlations across many pairs suddenly rise together, the market is often in a stress/risk-off regime.

⏰ Data & refresh schedule

Price data comes from Yahoo Finance via the yfinance library. The dashboard uses 91 trading days (~4.5 months) of daily closing prices, automatically excluding weekends and holidays via forward-fill.

Data is cached in Upstash Redis and refreshed nightly via a cron job, so repeated visits during the day are instant. To force a manual refresh, click the Refresh Data button in the header (useful after market close each trading day).