Pick a moment. The stone is the event; each ring is time since it happened.
A show, a film, a disaster, a discovery. Its ripples spread out in time, each one rated for how sure we are, until something sticks: a law, an institution, a change in how people live.
The stone is the event. A single wavefront leaves it and moves outward as time passes; each thing that changed appears where the front was when it happened, and stays there. So the inner arcs are the earliest changes and the outer ones the latest, and no ripple ever passes another. Time is stretched near the stone: the first weeks take as much room as the later decades, which is why the slider's ticks are uneven. Each distinct ripple gets its own slice, labeled on the rim with the industry it reached. Press Step and the clock stops at each ripple with its card highlighted; play or the right arrow goes on to the next.
Every step must begin after the step before it. A step that started first is busted, however related it looks.
Measured: moved beyond normal in the data, after the step before, and beat the same test run at fake dates in the series' own past. Most measured links are daily Wikipedia views, which show attention, not behavior; a few are official series (FRED, US baby names). Timed: rose in the right order, not tested against chance. Reported: a credible source makes the link. Plausible: it happened; the link is unproven. Busted: wrong order, contradicted, or no evidence. Steps without an exact date sit at the middle of their likely range. p is the share of placebo windows in a page's own history that rose as much as the real one; small means the rise is rare for that page, not that the step caused what came next. Every dated step since 2015 also gets an attention check: did Wikipedia views of the thing it names rise at that date, tested against the page's own past? It rides along on the card; it does not change the step's rating.
Ripple maps list things whose attention rose unusually after the event, found automatically among the Wikipedia articles that link to it. Attention is a lead, not a proven effect. The data: daily Wikipedia views, FRED and BLS series, Census retail sales, Zillow, Stack Overflow, US baby names and dated public records.
Ripple v1, an early build. Timing shows order, not proof of cause. Every test and every miss: ripple maps · discoveries. This page counts visits and plays anonymously with Google Analytics; no account, no ads.