Imagine we have data from 1 million app-based ride journeys in Seattle. Our goal is to develop a model that predicts the estimated time of arrival (ETA) once a ride request is made. How can we determine if our dataset is sufficient to build a model with reliable accuracy?