Users Segmentation
The leading urban mobility integration platform in Spain, Wondo, owned by Ferrovial, was facing a technological challenge following its recent merger with Moovit, the renowned mobility application, with more than 500,000 users in Madrid. With an enormous amount of data, including interactions and geolocation, the main obstacle was to improve the conversion rate of users, many of whom did not complete their registration with Wondo.
At WhiteBox, we undertook several initiatives to maximize the use of Moovit's extensive data for Wondo. Through a detailed analysis of user behavior, we identified critical points in the registration process and helped with their redesign to prevent them from abandoning the process halfway. We implemented an unsupervised clustering algorithm on geospatial data using open source libraries such as scikit-learn and Spark to distribute the computing, which allowed us to deduce the places of residence and work of the users and segment them by geographical areas of Madrid.
Our collaboration with WhiteBox brought significant benefits to Wondo:
- The new registration process increased user conversion from 50% to 80%.
- The geolocation algorithm revealed important patterns in the relationship between the location of users and their interaction with Wondo services, allowing efforts to be focused on the areas of greatest profitability.
“WhiteBox has helped us turn our data into actionable information. WhiteBox provides technical excellence throughout the data management process: extraction, processing, storage, and analysis. In our case, they also helped us set up and recruit our own team of data engineers.”