Since 2009, the social media Twitter provides free access to localization data consisting of geotagged messages sent on the platform. We used this data to highlight their potential contributions to the study of social groups’ activity space. Using a sample of luxury shoppers, we analyzed the daily activity and mobility patterns of the upper middle class in Delhi, India. We intersected these data with the results of an ethnographic study carried out in two luxury shopping places in Delhi and found that the lower middle class was overrepresented in the Twitter sample. We argue that the varying motivation to post on social media tend to skew the representativeness of social groups sampled on a spatial criterion. We eventually advocate the complementary use of geotagged social media data and qualitative data based on ethnographic research.