Real output from the engine — not a mockup. Every score is computed from 120 days of history with the math published in the methodology.
| Topic | Level | Score | Growth 14d | Signal |
|---|---|---|---|---|
| ChatGPT AI | rising | 58.4 | +30% | all sources agree |
| Wordle Fads | steady | 45.1 | +20% | 50% agreement |
| Temu Shopping | steady | 43.5 | +20% | 50% agreement |
| Air fryer Home | steady | 41.5 | +3% | 50% agreement |
| Berberine Health | steady | 40.9 | +0% | 50% agreement |
| Pilates Fitness | steady | 39.7 | −3% | 0% agreement |
↑ live data. Full report with charts: open the weekly report.
Deterministic scoring and honest data health — you always know where every number came from.
Wikipedia attention, global news volume (GDELT), and Google search-suggestion demand. All free, keyless, politely rate-limited.
Day-of-week seasonality removed. Gaps detected. Failed pulls are logged and shown, never silently dropped.
Growth vs prior window, momentum slope, breakout z-score, cross-source agreement — and a spike-decay fad penalty.