Combining Text Mining and Data Visualization Techniques to Understand Consumer Experiences of Electronic Cigarettes and Hookah in Online Forums

Introduction Since their introduction to the US market in 2007, electronic cigarettes (e-cigarettes) have posed considerable challenges to both public health authorities and government regulators, especially given the debate – in both the scientific world and the community at large – regarding the potential advantages (e.g. helping individuals quit smoking) and disadvantages (e.g. renormalizing smoking) associated with the product1. Similarly, hookah – a kind of waterpipe used to smoke flavored tobacco – has increased in popularity in recent years, is known to be particularly popular among younger people, and has prompted a range of regulatory responses2. One important – and currently largely unexplored – area of research involves exploring consumer perceptions and experiences of these emerging tobacco products. In this work, we use online health discussion forums in conjunction with text mining and novel data visualization techniques to investigate consumer perceptions and experiences of e-cigarettes and hookah, focusing on the automatic identification of symptoms associated with each product, and consumer motivations for product use. Previous related research has focused on using text-mining to analyze e-cigarette or hookah related Twitter posts3,4 and on the qualitative identification of e-cigarette related symptoms from online discussion forums5. The research reported in this abstract is – to the best of our knowledge – the first time that text mining techniques have been used with online health forums to understand e-cigarette or hookah use.