At Peregrine, we bring cameras into cars to gain awareness of the surrounding world and to create a superior data set for advanced mobility analytics. Video data captured in public spaces contain a lot of personal information (e.g. faces, license plates) which needs to be handled with special care. This talk will highlight how we address this challenge: From meta-data extraction and discarding of images directly at the edge, anonymization of the video material to intelligent data abstractions. And all in such a way, that we still can make use of the data for our analytics and model training.