This paper sets out to interpolate and predict state-level polling at the daily level by employing a dataset of over 500GB of political tweets from the final months of the 2012 presidential campaign.
This paper evaluates two tools for the automated identification of demographic and occupational data from the profile descriptions of Twitter users in the United Kingdom. Using human validation, the validity and reliability of automated machine processes is assessed for demographics and occupation. The authors find that it is possible to detect signatures of both occupation and age from Twitter metadata with varying degrees of accuracy, with less reliability towards occupation groups, but that further confirmatory work is required. These findings illustrate an incremental step forwards on the repurposing of social media data for social scientific analysis.
This paper is a useful starting point for researchers attempting to understand occupational and demographic-related data from extremist users on Twitter and other social media platforms.