Bottom line up front: what spiked, where the next campaign is forming, and which stories the data can already support — read for decisions, not decoration.
The signals worth knowing first — synthesized from the forecast and early-warning models.
Charges and locations whose arrest volume is spiking above their recent baseline — leading indicators of an emerging campaign. Ranked by statistical anomaly (z-score).
A z-score of 2+ means volume is well above the trailing baseline — worth a look, not a verdict. For coordinated multi-region campaign scoring, see the Emerging campaigns board below.
Prosecution grounds whose recent arrest volume runs well above their own baseline. Campaigns spanning multiple regions score higher — one city can be a zealous prosecutor; six regions is coordination.
A surge fires when recent arrests reach at least twice the volume the trailing baseline predicts, with a minimum-case floor. Case-basis enrichment lags fresh scrapes by a few weeks, so the newest campaigns understate before they overstate.
Platforms cited in case files as the venue of the alleged offense — a post, a donation, a repost — ranked by documented cases and the sentence-years attached to them.
This measures where prosecutions say the offense happened, not platform cooperation with authorities. High counts partly reflect platform popularity — read alongside the per-charge mix, not as a standalone ranking of danger.
Prophet time-series projection of arrest volume, with an 80% confidence band. The shaded region right of the marker is the forecast; everything left is the model fit to observed history.
Forecasts are statistical projections from historical arrest records and carry uncertainty — read the band, not the line. The upgraded model is event-aware (mobilization, elections, wartime censorship laws) with an 80% band conformal-calibrated against held-out backtests. Selections with sparse data fall back to the all-jurisdictions model.