Seasonality and relative-performance dashboard

Seasonality analysis is easy to do badly. Run enough windows against enough assets and something will always look significant, so the interesting engineering is in making the significance test explicit and adjustable rather than hiding it behind a chart. This dashboard decomposes each series into trend, seasonal and residual components and then applies configurable effect-size and consistency thresholds, so the user decides what counts as a real pattern instead of taking the tool's word for it.

Alongside it sits multi-asset relative-performance comparison over arbitrary windows, and the whole thing refreshes daily against public market and Federal Reserve data.