The engine found the COVID cliff in NYC taxi data without being told to look
The question we asked
> assuming today is 14 March 2020, run a trend and spike analysis on trip volume
What the engine found
No sustained trend across the period — and one sharp single-day drop on 14 March 2020, concentrated in the mid-fare, non-tipped segment.
Observed volume fell to 1,428 against a forecast of 2,114, well outside the confidence interval. The engine flagged it as worth investigating for an external disruption.
What it means
We know exactly what happened in New York on that date. The engine did not — it has no calendar of world events and no labels. It monitored every meaningful segment against its own forecast and surfaced the one that broke. That is the whole proposition: the finding did not depend on anyone suspecting there was something to find.
The honest caveat, and it belongs here: a single-day drop of this size is not subtle, and a competent analyst watching that chart would have seen it too. What they would not have done is check all 8,400 segments to establish that this one was anomalous and the others were not.
“Observed 1,428 against a forecast of 2,114 — well outside the confidence interval, on a day nobody had flagged.”
Check it yourself
NYC TLC trip record data is public. Dataset, dimensions and the exact window are listed to the right; the run takes about four minutes on commodity hardware.
Ask your first why.
Bring one question you've been putting off. We'll build the cube and answer it on the call.