Figure 2From: From individual to collective behaviours: exploring population heterogeneity of human mobility based on social media dataCharacteristics of geotagged activity of Swedish users. (A) Geotagged activity Origin-Destination Matrix (ODM) on the map. Each point represents a region formulated by DBSCAN clustering with the distance threshold for merging as 1 km and the minimum number of location for a region is set as 1 [10, 49]. (B) Geotagged activity ODM on the map where the trip has both origin and destination located within Sweden. (C) A week-long geotagging activity pattern that captures the time-dependent characteristic of geotagged locations (average of all the users). The warmer the colour (e.g. red and orange), the higher number of geotagged locations. (D) The distribution of the ratio of distinct geotagged locations over total geotagged locations (individually calculated). (E) Daily distributions of visiting frequency of the top two most visited locations, weekday vs. weekend (adjusted by the overall distribution of geotagged tweeting frequency over seven days across a week)Back to article page