Carpooling has the potential to be a key component of the transport mix in post-carbon, ecologically sustainable societies. To achieve this potential, it is imperative that the carpooling is transformed from an individualistic, private mode of transport to a hybrid private-public mode. Door-to-door matches are highly convenient for individual, sporadic carpool journeys , but hinder the development of carpooling as a mass transit service. The focus shifts to matching trajectory segments which are highly frequented by both passengers and drivers. Whilst this leads to a loss of personal convenience, as drivers and passengers are constrained to converge at meeting points along the trajectory segments, these meeting points act as aggregators to reach a critical mass of matched passenger demand and driver supply. In a world first, Ecov provides innovative carpooling services where passengers make carpooling requests without a priori matched drivers, and that these requests are stochastically matched in real-time to the driver flow on the pre-selected road segments. The mathematical complexity of stochastic carpooling matching greatly exceeds that of the traditional deterministic services. To address this complexity, we introduce a workflow, comprising of a combination of data science and GIS (Geographic Information Systems), which processes driver GPS traces in order to provide important indicators (e.g. driver participation rate and passenger waiting time) to guide the operational decision-making. We illustrate this workflow on a currently operational carpooling service in the peri-urban region surrounding the city of Lyon, France.