Dockless e-scooter sharing, as a new shared micromobility service, has quickly gained popularity in recent years. In this paper, we present a practical approach to estimating e-scooter flow patterns without knowing the actual routes taken by the e-scooter riders. Our method takes advantage of a huge open dataset that contains the origins and destinations of millions of trips. We show that our models can help cities better support the emerging shared micromobility service. The additional information generated in the modeling process can also be useful for a more refined analysis of e-scooter trips.