User profiling with geo-located posts and demographic data

Abstract

This paper presents a novel method for user profiling in social media that makes use of geo-location information associated with social media posts to avoid the need for selfreported data. These posts are combined with two publicly available sources of demographic information to automatically create data sets in which posts are labelled with socio-economic status. The data sets are linked by identifying each user's 'home location'. Analysis indicates that the nature of the demographic information is an important factor in performance of this approach.