More Like this

1. A deep semantic matching approach for identifying relevant messages for social media analysis

   https://doi.org/10.1038/s41598-023-38761-y  

   Biggers, Frederick Brown; Mohanty, Somya D.; Manda, Prashanti (July 2023, Scientific Reports)

   Abstract There is a growing interest in using social media content for Natural Language Processing applications. However, it is not easy to computationally identify the most relevant set of tweets related to any specific event. Challenging semantics coupled with different ways for using natural language in social media make it difficult for retrieving the most relevant set of data from any social media outlet. This paper seeks to demonstrate a way to present the changing semantics of Twitter within the context of a crisis event, specifically tweets during Hurricane Irma. These methods can be used to identify the most relevant corpus of text for analysis in relevance to a specific incident such as a hurricane. Using an implementation of the Word2Vec method of Neural Network training mechanisms to create Word Embeddings, this paper will: discuss how the relative meaning of words changes as events unfold; present a mechanism for scoring tweets based upon dynamic, relative context relatedness; and show that similarity between words is not necessarily static. We present different methods for training the vector model in Word2Vec for identification of the most relevant tweets for any search query. The impact of tuning parameters such as Word Window Size, Minimum Word Frequency, Hidden Layer Dimensionality, and Negative Sampling on model performance was explored. The window containing the local maximum for AU_ROC for each parameter serves as a guide for other studies using the methods presented here for social media data analysis. 

   more » « less
2. Clearing the air: evaluating institutions’ social media health messaging on wildfire and smoke risks in the US Pacific Northwest

   https://doi.org/10.1186/s12889-024-17907-1  

   Slavik, Catherine E; Chapman, Daniel A; Cohen, Alex Segrè; Bendefaa, Nahla; Peters, Ellen (December 2024, BMC Public Health)

   Abstract BackgroundWildfire smoke contributes substantially to the global disease burden and is a major cause of air pollution in the US states of Oregon and Washington. Climate change is expected to bring more wildfires to this region. Social media is a popular platform for health promotion and a need exists for effective communication about smoke risks and mitigation measures to educate citizens and safeguard public health. MethodsUsing a sample of 1,287 Tweets from 2022, we aimed to analyze temporal Tweeting patterns in relation to potential smoke exposure and evaluate and compare institutions’ use of social media communication best practices which include (i) encouraging adoption of smoke-protective actions; (ii) leveraging numeric, verbal, and Air Quality Index risk information; and (iii) promoting community-building. Tweets were characterized using keyword searches and the Linguistic Inquiry and Word Count (LIWC) software. Descriptive and inferential statistics were carried out. Results44% of Tweets in our sample were authored between January-August 2022, prior to peak wildfire smoke levels, whereas 54% of Tweets were authored during the two-month peak in smoke (September-October). Institutional accounts used Twitter (or X) to encourage the adoption of smoke-related protective actions (82% of Tweets), more than they used it to disseminate wildfire smoke risk information (25%) or promote community-building (47%). Only 10% of Tweets discussed populations vulnerable to wildfire smoke health effects, and 14% mentioned smoke mitigation measures. Tweets from Washington-based accounts used significantly more verbal and numeric risk information to discuss wildfire smoke than Oregon-based accounts (p = 0.042 andp = 0.003, respectively); however, Tweets from Oregon-based accounts on average contained a higher percentage of words associated with community-building language (p < 0.001). ConclusionsThis research provides practical recommendations for public health practitioners and researchers communicating wildfire smoke risks on social media. As exposures to wildfire smoke rise due to climate change, reducing the environmental disease burden requires health officials to leverage popular communication platforms, distribute necessary health-related messaging rapidly, and get the message right. Timely, evidence-based, and theory-driven messaging is critical for educating and empowering individuals to make informed decisions about protecting themselves from harmful exposures. Thus, proactive and sustained communications about wildfire smoke should be prioritized even during wildfire “off-seasons.” 

   more » « less
3. Toward Win-win Message Strategies: The Effects of Persuasive Message Content on Retweet Counts During Natural Hazard Events

   https://doi.org/10.1175/WCAS-D-20-0039.1  

   Li, Yajie; Hughes, Amanda Lee; Howe, Peter D. (March 2021, Weather, Climate, and Society)

   null (Ed.)

   Abstract Message diffusion and message persuasion are two important aspects of success for official risk messages about hazards. Message diffusion enables more people to receive lifesaving messages, and message persuasion motivates them to take protective actions. This study helps to identify win-win message strategies by investigating how an under-examined factor, message content that is theoretically important to message persuasion, influences message diffusion for official risk messages about heat hazards on Twitter. Using multilevel negative binomial regression models, the respective and cumulative effects of four persuasive message factors, hazard intensity, health risk susceptibility, health impact , and response instruction on retweet counts were analyzed using a dataset of heat-related tweets issued by U.S. National Weather Service accounts. Two subsets of heat-related tweets were also analyzed: 1) heat warning tweets about current or anticipated extreme heat events and 2) tweets about non-extreme heat events. This study found that heat-related tweets that mentioned more types of persuasive message factors were retweeted more frequently, and so were two subtypes of heat-related tweets. Mentions of hazard intensity also consistently predicted increased retweet counts. Mentions of health impacts positively influenced message diffusion for heat-related tweets and tweets about non-extreme heat events. Mentions of health risk susceptibility and response instructions positively predicted retweet counts for tweets about non-extreme heat events and tweets about official extreme heat warnings respectively. In the context of natural hazards, this research informs practitioners with evidence-based message strategies to increase message diffusion on social media. Such strategies also have the potential to improve message persuasion. 

   more » « less
4. Analysis of lockdown perception in the United States during the COVID-19 pandemic

   https://doi.org/10.1140/epjs/s11734-021-00265-z  

   Surano, Francesco_Vincenzo; Porfiri, Maurizio; Rizzo, Alessandro (September 2021, The European Physical Journal Special Topics)

   Abstract Containment measures have been applied throughout the world to halt the COVID-19 pandemic. In the United States, several forms of lockdown have been adopted in different parts of the country, leading to heterogeneous epidemiological, social, and economic effects. Here, we present a spatio-temporal analysis of a Twitter dataset comprising 1.3 million geo-localized Tweets about lockdown, from January to May 2020. Through sentiment analysis, we classified Tweets as expressing positive or negative emotions about lockdown, demonstrating a change in perception during the course of the pandemic modulated by socio-economic factors. A transfer entropy analysis of the time series of Tweets unveiled that the emotions in different parts of the country did not evolve independently. Rather, they were mediated by spatial interactions, which were also related to socio-ecomomic factors and, arguably, to political orientations. This study constitutes a first, necessary step toward isolating the mechanisms underlying the acceptance of public health interventions from highly resolved online datasets. 

   more » « less
5. Understanding underlying moral values and language use of COVID-19 vaccine attitudes on twitter

   Judith Borghouts, Yicong Huang (January 2023, PNAS nexus)

   Public sentiment toward the COVID-19 vaccine as expressed on social media can interfere with communication by public health agencies on the importance of getting vaccinated. We investigated Twitter data to understand differences in sentiment, moral values, and language use between political ideologies on the COVID-19 vaccine. We estimated political ideology, conducted a sentiment analysis, and guided by the tenets of moral foundations theory (MFT), we analyzed 262,267 English language tweets from the United States containing COVID-19 vaccine-related keywords between May 2020 and October 2021. We applied the Moral Foundations Dictionary and used topic modeling and Word2Vec to understand moral values and the context of words central to the discussion of the vaccine debate. A quadratic trend showed that extreme ideologies of both Liberals and Conservatives expressed a higher negative sentiment than Moderates, with Conservatives expressing more negative sentiment than Liberals. Compared to Conservative tweets, we found the expression of Liberal tweets to be rooted in a wider set of moral values, associated with moral foundations of care (getting the vaccine for protection), fairness (having access to the vaccine), liberty (related to the vaccine mandate), and authority (trusting the vaccine mandate imposed by the government). Conservative tweets were found to be associated with harm (around safety of the vaccine) and oppression (around the government mandate). Furthermore, political ideology was associated with the expression of different meanings for the same words, e.g. “science” and “death.” Our results inform public health outreach communication strategies to best tailor vaccine information to different groups. 

   more » « less