How do you identify operational bottlenecks in a dataset tracking driver onboarding from registration to their first completed trip on Uber Drivers?
