Regional insights on tobacco-related tweets: unveiling user opinions and usage patterns

Abstract

Introduction: Tobacco consumption and its impact on health remain high worldwide. Additionally, it is a contentious issue generating significant controversy. Twitter has proven to be a useful platform for evaluating public health topics related to population health behaviors, and tobacco consumption.

Objective: The objective of this study is to analyze the content of tweets related to tobacco. Moreover, geolocation data will be considered to understand regional differences.

Methods: Tweets published between 2018 and 2022, in both English and Spanish, containing the keyword "tobacco," were analyzed. A total of 56,926 tweets were obtained. The tweets were classified into different categories. 550 tweets were manually analyzed, and an automated and computerized classification was performed for the remaining and largest subset of tweets.

Results: The analysis yielded 30,812 classifiable tweets. Healthcare professionals were the most frequent contributors to the topic (50.2%), with the most common theme being general information about the toxic effects of tobacco. 57.9% of the tweets discussed the harmful effects of tobacco on health, with fear being the predominant emotion. The largest number of tweets were located in America.

Conclusions: Our study revealed a substantial number of tweets highlighting the health risks and negative perceptions of tobacco consumption. Africa showed the lowest percentage of tweets discussing the health risks associated with tobacco, coinciding with the continent having the least developed anti-tobacco policies. Healthcare professionals emerged as the most prominent users discussing the topic, which is encouraging as they play a crucial role in disseminating accurate and scientific health information.

Keywords: Machine Learning; Twitter; artificial intelligence; geolocation; perception; tobacco.