On the Accuracy of Hyper-local Geotagging of Social Media Content

TL;DR

This paper introduces a data-driven method to estimate hyper-local geographic locations for social media content that lacks explicit geotags, analyzing trade-offs and platform differences to improve geolocation accuracy.

Contribution

It presents a novel approach modeling hyper-local n-gram distributions to predict locations of non-geotagged social media posts, considering platform and device variations.

Findings

Content from different sources shows distinct geographic patterns.

Modeling source-specific distributions improves location prediction.

Trade-offs between accuracy, coverage, and precision are characterized.

Abstract

Social media users share billions of items per year, only a small fraction of which is geotagged. We present a data- driven approach for identifying non-geotagged content items that can be associated with a hyper-local geographic area by modeling the location distributions of hyper-local n-grams that appear in the text. We explore the trade-off between accuracy, precision and coverage of this method. Further, we explore differences across content received from multiple platforms and devices, and show, for example, that content shared via different sources and applications produces significantly different geographic distributions, and that it is best to model and predict location for items according to their source. Our findings show the potential and the bounds of a data-driven approach to geotag short social media texts, and offer implications for all applications that use data-driven approaches to locate content.