Data Analysis and Visualization of Newspaper Articles on Thirdhand Smoke: A Topic Modeling Approach

Abstract

Background Thirdhand smoke has been a growing topic for years in China. Thirdhand smoke (THS) consists of residual tobacco smoke pollutants that remain on surfaces and in dust. These pollutants are re-emitted as a gas or react with oxidants and other compounds in the environment to yield secondary pollutants. Objective Collecting media reports on THS from major media outlets and analyzing this subject using topic modeling can facilitate a better understanding of the role that the media plays in communicating this health issue to the public. Methods The data were retrieved from the Wiser and Factiva news databases. A preliminary investigation focused on articles dated between January 1, 2013, and December 31, 2017. Use of Latent Dirichlet Allocation yielded the top 10 topics about THS. The use of the modified LDAvis tool enabled an overall view of the topic model, which visualizes different topics as circles. Multidimensional scaling was used to represent the intertopic distances on a two-dimensional plane. Results We found 745 articles dated between January 1, 2013, and December 31, 2017. The United States ranked first in terms of publications (152 articles on THS from 2013-2017). We found 279 news reports about THS from the Chinese media over the same period and 363 news reports from the United States. Given our analysis of the percentage of news related to THS in China, Topic 1 (Cancer) was the most popular among the topics and was mentioned in 31.9% of all news stories. Topic 2 (Control of quitting smoking) was related to roughly 15% of news items on THS. Conclusions Data analysis and the visualization of news articles can generate useful information. Our study shows that topic modeling can offer insights into understanding news reports related to THS. This analysis of media trends indicated that related diseases, air and particulate matter (PM 2.5 ), and control and restrictions are the major concerns of the Chinese media reporting on THS. The Chinese press still needs to consider fuller reports on THS based on scientific evidence and with less focus on sensational headlines. We recommend that additional studies be conducted related to sentiment analysis of news data to verify and measure the influence of THS-related topics.

Characterizing Artificial Intelligence Applications in Cancer Research: A Latent Dirichlet Allocation Analysis

“…Classifying information sources and content in corresponding topics to identify priorities for interventions has been widely applied in many studies. For example, previous authors have analyzed newspaper and social media content to understand topics of interest related to breast cancer and secondhand smoking [23-28]. However, none of the previous studies have analyzed the scientific bibliography to determine the development of research landscapes in AI applied in cancer care.…”

Categorising patient concerns using natural language processing techniques

“…Knowledge that two posts have come from the same discussion thread constitutes valuable domain information that can inform topic modelling, and LabEL is perfectly suited for this kind of analysis. As another example, Liu et al 33 analyse the topics present in articles about thirdhand smoke extracted from new databases, including Factiva. 50 Factiva uses a proprietary taxonomy called Dow Jones Intelligent Identifiers to provide labels to articles, including data elements such as region, topic and company (among others).…”

“…14 One particular unsupervised topic modelling method, latent Dirichlet allocation (LDA), 15 has proven particularly popular and successful. LDA has been used for topic mining in studies of health data across an array of data sources, including discussions from condition-specific online support groups [16][17][18][19][20] and more general online discussion platforms, [21][22][23][24][25][26][27][28][29] data about adverse medical events, 30 interview transcripts of patients, 31 32 media articles 33 and survey data. 34 35 Other studies have used LDA to analyse topics in patient-reported concerns as well, in situations where no existing topic information is available.…”

Health Communication through Chinese Media on E-Cigarette: A Topic Modeling Approach

“…The reason why we chose LDA topic modeling is because it is a type of topic modeling method that has been utilized in a number of studies for text mining [30], or in fields of psychology and medicine [31,32], as well as our team's previous research on third-hand smoke [33]. This unsupervised machine learning technique automatically generates topics from documents and categorizes similar documents to one or more of these topics based on the distribution of words, so we can implement a lot of text analysis.…”