Case Study: Parkeon predicts parking availability and reduces search time with Dataiku

How Parkeon Built a B2C Data-Driven Parking Prediction Application for Android and iPhone

Parkeon, a global supplier of parking and transit payment systems found in over 3,500 cities, needed to turn large volumes of parking-meter data into a reliable B2C parking-availability prediction app. To build a simple, dependable user experience that enriches meter data with geographic context, Parkeon partnered with Dataiku and used Dataiku’s Data Science Studio (DSS).

Using Dataiku’s DSS, Parkeon combined millions of parking transactions with OpenStreetMap points of interest, segmented streets, and trained clustering and mixture models to power the iOS app “Path to Park.” The solution predicts three nearby parking options with more than an 80% probability of finding a spot, enables fast iteration and scalability across cities, and embeds the predictive algorithm directly in the app to improve driver routing and parking success.

Parkeon

Yves-Marie Pondaven

CTO