Sentiment analysis on social media for identifying public awareness of type 2 diabetes

Abstract

Many studies have investigated how Type 2 Diabetes Mellitus (T2DM) is impacted by environmental characteristics. However, it is unclear whether the public is aware of this relationship. This paper aims to infer public awareness based on sentiment analysis of environmental characteristics from social media (news) as context-based clues for supporting planners and designers to achieve a health-supportive environment. We apply natural language processing (NLP) techniques to past years’ news (from 1992 to 2020) to understand the public awareness on the relationship between the prevalence of T2DM and environmental characteristics. The results not only verify that the public aware little that prevalence of T2DM correlates with environmental characteristics in Singapore (1% news related to both diabetes and environment), but also present a hypothetical urban design study to demonstrate how this approach can be adapted to identify the health-related urban issues at the early design stage. This study can offer urban planners and designers an innovative design perspective to incorporate residents' health into the design process.