One-way weekend trips dominate bad outcomes in ride-hailing
The question we asked
> which trip segments are driving poor outcomes across the network?
What the engine found
Five separate analyses — decomposition, anomaly detection, pattern mining, drift and a predictive model — were run against the same trip data. One-way trips taken on weekends topped the ranking in every one of them, not just the analysis it was best suited to.
The same segment led all five independently run analyses, which is a stronger signal than any single one on its own.
What it means
Agreement across five different methods, run independently rather than tuned to find the same answer, is what makes this segment worth acting on rather than a false positive from one algorithm's blind spot.
“The same segment led all five independently run analyses.”
Ask your first why.
Bring one question you've been putting off. We'll build the cube and answer it on the call.