Andria Systems · Quantitative Research

Institutional Equity
Intelligence Platform

Processes 116 million SEC 13F filings to surface regime-conditioned activist conviction signals. Combines unsupervised manager clustering with Gaussian Hidden Markov Model macro detection to rank institutional equity plays.

Current Macro Regime
Goldilocks
87% HMM confidence

Why These Numbers Matter

Three metrics that tell you whether the research is trustworthy before you look at a single signal.

98.5%
Data Provenance Quality

Share of holdings records that matched a live exchange ticker via OpenFIGI. High provenance means signals are based on real, tradeable securities, not stale or delisted names.

1.847
Out-of-Sample Sharpe Ratio

Annualised risk-adjusted return across 10 walk-forward folds, 2010–2024. Each fold trains on expanding history and tests on a held-out year with no hindsight bias.

87%
Model Confidence · Goldilocks

Probability assigned by the Gaussian HMM to the current macro regime state. Drives the regime multiplier applied to every RACS score. The higher the confidence, the stronger the signal tilt.

How the Pipeline Works

Six stages from raw SEC filings to validated, deployable signals. Click any stage to understand what it does, how it works, and why it matters.

Click any stage to expandPipeline v4.16 · DuckDB 0.10 · Polars 0.20

Pipeline Run History

Each successful run commits fresh artifacts to the repository, rebuilding this dashboard automatically via CI/CD.

20260525T183200_a4f8e1
5/25/2026, 18:32:00 UTC
14m 22s
SUCCESS
20260524T183000_c9d2b3
2026-05-24 18:30:00 UTC
15m 01s
SUCCESS
20260522T183000_f1e8a2
2026-05-22 18:30:00 UTC
02m 14s
FAILED