Fig. 4
From: Deviations from typical paths: a novel approach to working with GPS data in the behavioral sciences
Differential separation of deviations on commute days with and without shopping across competing summary measures. For both the estimated typical path and the true commute route, using the average of all calculated deviations from a typical path provides the greatest separation between two similar types of days—a day in which the person only commutes and a day in which they commute and shop—while the traditional measure using a single distance referenced to a single point offers minimal separation. Separation is greater when using the true commute route as a reference rather than the estimated typical path. In this comparison, Haversine distances in meters are used for all measures but the range of values displayed differs across the graphs. The emphasis is on the differing separation between day types within each graph, rather than comparison of each type of day across graphs.