📈 Market-intelligence signals
What this is & how to consume it
A GDELT-style sentiment + buzz feed over every company, ticker, crypto and commodity mentioned across the scraped boards. For each entity we compute, per time-window:
- tone (−100…+100) — VADER sentiment of the discussion, mapped to a bullish/bearish axis.
- mention volume + a z-score vs the entity's own baseline (abnormal-attention signal).
- tone momentum — is the mood improving or souring.
- tone dispersion — how contested the mood is (high = a fight).
- buzz_index (0–100) and a composite signal label, with a confidence.
⚠️ This is anonymous imageboard chatter — noisy, manipulable, and
NOT financial advice. Treat it as one weak, contrarian-leaning sentiment input,
weighted by the confidence and tone_dispersion fields.
🟢 Most bullish (positive tone)
| Entity | Cat | Tone | Buzz | n | src |
|---|
🔴 Most bearish (negative tone)
| Entity | Cat | Tone | Buzz | n | src |
|---|
🔥 Highest buzz (composite heat)
| Entity | Cat | Buzz | Signal | z | src |
|---|
📡 Volume spikes (abnormal attention)
| Entity | Cat | z-score | Tone | Signal | src |
|---|
🔎 Inspect an entity timeline
📊 Mentions over time
Mention count per time bucket (bars). Spikes = abnormal attention.
📈 Sentiment over time
Average tone per time bucket (−100 bearish … +100 bullish); the dashed line is the neutral zero baseline.
🔌 API — for trading bots / notebooks
Stable JSON, no auth, CORS-friendly. window accepts 24h, 7d, 30d or any <N>h/<N>d.
| Endpoint | What you get |
|---|---|
GET /api/v1/signals?window=24h&category=crypto | Ranked scan of all market entities (lean snapshot, no series). |
GET /api/v1/signals/summary?window=24h | Leaderboards: most bullish / bearish / buzz / volume-spikes. |
GET /api/v1/signals/{entity}?window=7d | One entity, full timeline (volume_series + tone_series). |
Each signal object carries: tone_avg, mention_zscore, tone_momentum,
tone_dispersion, buzz_index, signal, confidence, boards.
Example: /api/v1/signals/bitcoin?window=30d