Mapping obesity and diabetes' representation on Twitter: the case of Italy.

Abstract

One of the main functions of public health is to monitor population health to identify health problems and priorities. Social media is increasingly being used to promote it. This study aims to investigate the field of diabetes and obesity and related tweets in the context of health and disease. The database extracted using academic APIs (Application Programming Interfaces) allowed the study to be run with content analysis and sentiment analysis techniques. These two analysis techniques are some of the tools of choice for the intended objectives. Content analysis facilitated the representation of a concept and a connection between two or more concepts, such as diabetes and obesity, on a purely text-based social platform such as Twitter. Sentiment analysis therefore allowed us to explore the emotional aspect related to the collected data related to the representation of such concepts. The results show a variety of representations connected to the two concepts and their correlations. From them it was possible to produce some clusters of elementary contexts and structure narrative and representational dimensions of the investigated concepts. The use of sentiment analysis and content analysis and cluster output to represent complex contexts such as diabetes and obesity for a social media community could increase knowledge of how virtual platforms impact fragile categories, facilitating concrete spillovers into public health strategies.