Asa a big fan of event sourcing, it was nice to see a well considered and accessible article about it. Good job.

The particular example he used was poor. GPS coordinates along the way of a ride are meaningful for ridesharing, because you need to know the route the driver took along the entire route from start to end. You also need to know where the driver is prior to a ride being started because you need to match a driver to a passenger, and location matters a lot for that. You also probably want to know the route updates during a ride so that you can match the driver against more passengers mid-ride (either for carpooling, or for chained rides).

Ideally you'd use both the passenger and driver data to ensure the driver's client isn't misleading you, in case you're paying by time and distance, and also to ensure the passenger's safety (are they actually taking the passenger somewhere else?).

Local law may require you to have a particular granularity for time/distance calculations.

The article's point still generally stands, though. You need to decide for yourself what data you're sending, and the frequency of what you're sending, and you're using the raw data to do so. How you choose this should be based on business need, not directly from raw data being produced.