What is Stratum

Stratum is a quantitative data-lake platform. It turns raw external data — macroeconomic series, market data, regulatory filings — into refined, provenance-tracked quantitative signals.

The shape of it

• Bronze — raw source responses, stored verbatim.
• Silver — normalized, deduplicated, schema-validated Parquet.
• Gold — signals: cross-source joins, scored and aggregated for analysis.

In these docs

The pages under Data Lake are published from the live stratum-datalake project documentation:

• Architecture — how the platform is structured and deployed.
• Gold signals — the signals Stratum produces and how to read them.
• Research layer — the validation methodology behind the signals.
• Cross-source joins — how data from different sources is combined.

Note

These docs are synced from the engineering repository at build time. Content that is internal-only (operations, compliance, security baselines) is excluded by an explicit allowlist.