Geolocation for Twitter: Timing Matters

Proceedings of NAACL-HLT 2016, pages 1064–1069, San Diego, California, June 12-17, 2016. c⃝2016 Association for Computational Linguistics Geolocation for Twitter: Timing Matters Mark Dredze1,2, Miles Osborne1, Prabhanjan Kambadur1 1 Bloomberg L.P. 731 Lexington Ave, New York, NY 10022 2 Human Language Technology Center of Excellence Johns Hopkins University, Baltimore, MD 21211 mdredze@cs.jhu.edu mosborne29,pkambadur@bloomberg.net Abstract Automated geolocation of social media mes- sages can benefit a variety of downstream ap- plications. However, these geolocation sys- tems are typically evaluated without attention to how changes in time impact geolocation. Since different people, in different locations write messages at different times, these factors can significantly vary the performance of a ge- olocation system over time. We demonstrate cyclical temporal effects on geolocation accu- racy in Twitter, as well as rapid drops as test data moves beyond the time period of training data. We show that temporal drift can effec- tively be countered with even modest online model updates. 1 Introduction Geolocation – the task of identifying a social media message’s location – can support a variety of down- stream applications, such as advertising, personal- ization, event discovery, trend analysis and disease tracking (Watanabe et al., 2011; Hong et al., 2012; Kulshrestha et al., 2012; Broniatowski et al., 2013). Geolocation work has mostly focused on Twitter, since tweets are readily accessible and true loca- tion available from user geocoded tweets (inter alia (Eisenstein et al., 2010; Han et al., 2014; Rout et al., 2013; Compton et al., 2014; Cha et al., 2015; Jur- gens et al., 2015; Osborne et al., 2014; Dredze et al., 2013)). Most previous work consider the task of author geolocation, the identification of a author’s primary (home) location (Eisenstein et al., 2010; Han et al., 2014). Author geolocation systems rely on multiple tweets from each author to identify the location. In this work, we consider the task of tweet geolocation, where a system identifies the location where a single tweet was written (Osborne et al., 2014; Dredze et al., 2013). This approach is necessary when geolo- cation decisions must be made quickly, with limited resources, or when the location of a specific tweet is required. When focusing on a single tweet, time becomes relevant. Intuitively, tweets written in the morning might be in different locations (at home) than say tweets written during the day (at work). This in- formation is often ignored but can provide impor- tant clues as to a tweet’s location. Likewise, mod- els built using historical data never adapt as time evolves. These factors may have a significant impact on geolocation accuracy, and downstream system’s should be sensitive to these variations. For the first time, we consider the impact of time on Twitter geolocation and predict where a post was made (rather than the more usual, and easier task of author location). We take a supervised learning approach, training a multi-class classifier to identify the city of a tweet. We train a system on 250 million tweets sampled from a 45 month period, perhaps the largest evaluation to …

Content truncated to 500 words