Transparent by design

Methodology & data boundaries

How SignalDesk turns heterogeneous event streams into explainable monitoring cues.

Normalize

Market, news, filing, company and forum records map to a shared ticker-keyed event envelope with source lineage.

Measure

Rolling volume and relative volume compare the current interval with session-adjusted historical baselines.

Correlate

Mention velocity, sentiment, price displacement and verified events remain independent features before scoring.

Verify

Official filings and company sources receive primary reliability. Licensed reporting is verified; social content remains social.

Rank

Discovery ranking is volume-first: sudden five-minute spikes carry 50%, sustained relative volume 20%, verified news 20% and sentiment strength 10%.

Evolve

Typed provider interfaces separate the educational model from future licensed market, news, SEC, Reddit and Stocktwits connectors.

Current discovery formula

Volume-first discovery score

50 × sudden 5m volume spike + 20 × relative volume + 20 × verified news strength + 10 × sentiment strength

Stocks are ranked independently inside each market-cap category. Volume contributes 70% of the score, so a sudden spike can surface before a news or sentiment narrative develops. The score is a monitoring cue, not a prediction.

Responsible-use boundary

This application provides educational market intelligence, not financial advice, an offer, or a personalized recommendation. Illustrative sources and URLs show the intended lineage pattern. Always validate facts with official filings and primary sources.