Geotagging matters? The interplay of space and place in politicized online social media networks

Voting results in marginal constituencies often determine wider political outcomes. It is now apparent that key electorates in these areas have been geo-behaviourally targeted by elaborate operations intended to manipulate results through advertising, (mis)information, and/or ‘fake news’ disseminated via online social networks. Attempts to track the geographical diffusion of cyber politicking are hindered by incomplete geospatial referencing in social media (meta)data. Just about 1–2% of publicly posted Twitter tweets, and even fewer Facebook posts, are typically ‘geotagged’ with Latitude and Longitude coordinates. Many more records (about 25%) make toponymic mention of place. This paper examines about 8 million social media interactions, over 350,000 of which are geotagged, created during the 2012 US Presidential Election and the 2014 Scottish Independence Referendum campaigns, to assess the interplay of space and place in online communications. Results of text and data-mining show that coordinate-geotagging users of Twitter and Facebook, (a) make fewer references to place in their message text, (b) link to articles making fewer mentions of place in their content, and (c) make far fewer links to external content than their non-coordinate-geotagging peers. Despite providing some valuable geospatial information, coordinate-geotagged interactions offer only an inadequate proxy for tracking the spread of all places, linked content, or (mis)information shared online. As Twitter retires its tweet spatialization functionality, new regulatory and technical responses together with a better understanding of place will be required if electoral officials, platform operators, and researchers are to more easily and accurately identify nefarious content targeting specific areas as well as specific individuals during democratic elections.