Identification and analysis of medical entity co-occurrences in twitter

Abstract

Twitter is an attractive source of data for public health surveillance, as it is less hindered by the legal and techni-cal obstacles associated with data sources such as electronic health records. We present a preliminary co-occurrence anal-ysis based on 10% of all tweets from 2014 annotated with medical entities as a first approach to extract health-related facts from Twitter. In this work, co-occurrence of anno-tated medical entities are used to provide population-scale information about common health issues and related enti-ties, which has potential applications in areas such as phar-macovigilance.