INTERACTIVE VISUALIZATION
> This page brings together four views of the same underlying data: the correlation network built from daily log-returns of the sample universe. Each chart shows a different facet of the network's geometry. Together they answer the question: where is the network under stress, and how is that stress evolving?
All charts are interactive — hover for values, drag to rotate the 3D graph, and click edges or cells to cross-highlight. The Controls panel at the bottom lets you filter by correlation threshold and toggle edge visibility.
CURVATURE TIME SERIES
Curvature Over Time
Shows mean network curvature history. The shaded area shows min/max range. Hover to explore values. Watch for sustained drops below zero.
Hover over chart to explore dates
Reading this chart:
The green line tracks the network's mean Ollivier-Ricci curvature over time. The shaded band shows the min-to-max range across all edges. When the line is above zero the network is geometrically "well-connected" — correlations are diversified. A sustained drop toward or below zero signals rising concentration and fragility. Dotted vertical lines mark known stress events for reference.
3D NETWORK GRAPH
(latest snapshot)
Correlation Network
3D visualization of asset correlations. Nodes are stocks, edges connect correlated pairs. Green edges = positive curvature (stable). Red edges = negative curvature (stressed). Shows most recent data.
Drag to rotate
Reading this chart:
Each node is a stock in the sample universe. Edges connect pairs whose correlation exceeds the threshold (default |ρ| > 0.3). Green edges have positive curvature — the two nodes share many common neighbors, so the connection is "well-supported." Red edges have negative curvature — a bridge-like connection with few alternative paths, making it a potential stress conduit. Drag to rotate; unlock zoom with the button above.
CORRELATION MATRIX
(latest snapshot)
Pairwise Correlations
Heatmap of return correlations between assets. Green = positive correlation, Red = negative. High correlations during stress indicate contagion risk. Shows most recent data.
Click cell to highlight edge
Reading this chart:
This heatmap shows the 60-day rolling Pearson correlation between every pair of stocks. Bright green cells indicate strong positive co-movement; red cells indicate negative correlation. During market stress, off-diagonal cells tend to turn uniformly green as diversification breaks down — this is exactly the kind of regime shift that curvature is designed to detect. Click any cell to highlight the corresponding edge in the risk table below.
HIGHEST RISK EDGES
(latest snapshot)
Riskiest Connections
Asset pairs with most negative curvature. These edges act as "bottlenecks" in the network - stress here can propagate to the broader market. Shows most recent data.
Click bar to highlight edge
Reading this chart:
These are the edges with the most negative curvature in the latest snapshot — the network's structural weak points. A strongly negative bar means that pair of stocks forms a "bridge" between otherwise disconnected clusters. If the correlation on that bridge shifts suddenly, the shock has few alternative paths to dissipate. Think of these as the load-bearing beams of the financial network: they deserve the most monitoring.
CONTROLS
Correlation Threshold
0.30
Display Options
Color By
MORE VISUALIZATIONS