Use of geolocated tweets in human mobility research


Twitter’s terms of service says you can’t aggregate geo data associated with a tweet. The more reading I do about this part of the Terms of Service, the more confusion I have. Several studies (and online web sites) aggregate Twitter’s content AND geolocation to understand some aspect of humans (e.g., content word = “flu”). Other studies (and online papers) aggregate Twitter’s geolocation only (not content) to understand human movement; in a sense, these studies are using ALL content with geolocation (i.e., content word = ANY). In all these cases, the results would not surprise a user, nor violate a user’s privacy (as the aggregation is not done at an individual’s level). Thus, I am concluding that both these types of research are acceptable. I know official Twitter moderators are reading these posts. I will assume my conclusion is accurate if I do not hear back from Twitter in one week (April 25th, 2014). Thanks for reading!