Case Study: CKDelta predicts train station crowd density in real time with Quix

Predicting crowd density for public transport in real time

CKDelta worked with Quix to build a real-time event streaming application that could predict crowd density and waiting times at Singapore train stations. The challenge was to process Wi‑Fi data from 180 underground stations quickly and securely, while meeting a tight four-month deadline and supporting a public-facing mobile app to help reduce overcrowding.

Quix provided a production-ready, modular stream processing platform that let CKDelta’s data scientists deploy Python ML models without re-engineering, while the platform team integrated the system into existing infrastructure on a dedicated, locked-down AWS environment. The solution went live on time, handled nearly 40GB of data per day versus the planned 10GB, delivered 4x the anticipated processing capacity, and proved reliable enough for CKDelta to consider expanding it to more cities and locations.

CKDelta

Fernando Ayuso

Director of Data Science and Data Engineering