Using Machine Learning and Twitter Data to Evaluate the Public Opinion of the Delta Variant and the Vaccine Booster
The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.
By April Heyward October 17, 2021
Public administration has crossed over the 18 month mark of combating the COVID-19 pandemic with health, economic, education and social welfare public policy interventions. Traditional public service delivery was suspended in the early stages of the COVID-19 pandemic and public administration employed innovation to reinvent public service delivery, while simultaneously combating the COVID-19 pandemic. COVID-19 continues to be a shocking and sobering infectious disease and the virus responsible for COVID-19 has expanded with Variants of Concern, Variants of Interest and Variants Under Monitoring globally over the last several months. The World Health Organization (WHO) employs the Greek alphabet to characterize Variants of Concern (e.g., Alpha, Beta, Gamma, Delta) and Variants of Interest (e.g., Lambda, Mu). There is variation in variant dominance globally and the Delta Variant became dominant in the United States in summer 2021. The Delta Variant was characterized by WHO as a Variant of Interest on April 4, 2021 and was reclassified as a Variant of Concern on May 11, 2021. April Heyward has been examining COVID-19 and public policy interventions as part of her research, and employed machine learning and Twitter data to evaluate the public opinion of the Delta Variant and the Vaccine Booster. Public administration can employ computational methods such as machine learning to evaluate public opinion and policy interventions. This serves as a nod to computational social science.
Machine learning is a fascinating subdiscipline of artificial intelligence and its vastness facilities natural language processing to evaluate public policy and public opinions. Social media data is a treasure trove for data scientists, and coupled with machine learning, can extend the arsenal of public administration. Heyward started extracting Twitter data via RStudio (an Integrated Development Environment for the R Programming language) with the rtweet package developed by Michael Kearney to study the public opinions of COVID-19, including prevention and risk management measures, and public policy interventions in February 2021. Retweets are excluded from data extraction. Datasets are created with each extraction and are saved as .CSV files. The extraction of Delta Variant Twitter data began as the variant became dominant in the United States and the extraction of Vaccine Booster Twitter data began as formal deliberations of the availability of the booster shot commenced. The text of tweets is cleaned in RStudio by removing punctuations, http links, stopwords, digits, other symbols and capital letters which are transformed to lowercase letters for preparation of the emotions analysis. Lexicons are dictionaries that ascribe words to positive and negative sentiments and emotions. Employing multiple lexicons can yield varying results that can extend understanding of sentiments and emotions. The NRC lexicon, developed by Saif Mohammad and Peter Turney, and the Bing lexicon, developed by Bing Lui, were employed in the emotions analysis of the Delta Variant and Vaccine Booster tweets. There are numerous methods for evaluating tweets that are inclusive of, but not limited to, the analysis of the sources of tweets, positive and negative emotions found in tweets and words that contribute to positive and negative emotions.
The Delta Variant Twitter data employed in the emotions analysis was extracted on September 24, 2021 and yielded 17,918 tweets from the September 20-24, 2021 time period. A subsequent Delta Variant Twitter dataset was created retaining tweets resulting from Twitter for Android, Twitter for iPad, Twitter for iPhone, Twitter for Mac and Twitter Web App sources. This step reduced the number of tweets to 13,894. The Vaccine Booster Twitter data employed in the emotions analysis was extracted on the same day and yielded 17,615 tweets from the September 18-24, 2021 time period. A subsequent Vaccine Booster Twitter dataset was created, retaining tweets from the same Tweet sources and netted 12,424 tweets. See Figure 1 for the Delta Variant No Hashtag Tweets by Tweet Source Donut Chart. The Twitter Web App was the top source for Delta Variant tweets at 40.02% and then Twitter for iPhone (31.92%), Twitter for Android (23.76%), Twitter for iPad (4.18%) and Twitter for Mac (0.12%) respectively. See Figure 2 for Words Contributed to Positive and Negative Emotions Found in Delta Variant No Hashtag Tweets. Figure 2 depicts the emotions and words that contributed to the emotions found in Delta Variant tweets. The words found in Delta Variant tweets ascribed more to the positive emotion than the negative emotion when using the NRC lexicon. The word “vaccine” emerged as the most used and was ascribed to the positive emotion.
Figure 1
Figure 2
The vaccine booster is an intervention to continue combating COVID-19. See Figure 3 for Vaccine Booster No Hashtag Organic Tweets—Emotions Analysis of September 18-24, 2021 Tweets—NRC Lexicon. The positive emotion emerged as the leading emotion in the analysis as compared to other emotions. Employing the Bing lexicon to identify words contributing to positive and negative emotions found in Vaccine Booster tweets added more context to the results. See Figure 4 for Words Contributed to Positive and Negative Emotions Found in Vaccine Booster No Hashtag Tweets. The top ten words were identified in the Vaccine Booster tweets that contributed to positive and negative emotions. Machine learning is a method for public administration to adapt from the computer science discipline and to employ.
Figure 3
Figure 4
Author: April Heyward is an Author for PA TIMES, Public Sector Practitioner in South Carolina, 4th Year Doctor of Public Administration Student at Valdosta State University, and a R Programmer. She can be reached at [email protected] and followed on Twitter: https://twitter.com/heyward_april. All opinions and views are her own and does not reflect the views and opinions of her affiliations.
Using Machine Learning and Twitter Data to Evaluate the Public Opinion of the Delta Variant and the Vaccine Booster
The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.
By April Heyward
October 17, 2021
Public administration has crossed over the 18 month mark of combating the COVID-19 pandemic with health, economic, education and social welfare public policy interventions. Traditional public service delivery was suspended in the early stages of the COVID-19 pandemic and public administration employed innovation to reinvent public service delivery, while simultaneously combating the COVID-19 pandemic. COVID-19 continues to be a shocking and sobering infectious disease and the virus responsible for COVID-19 has expanded with Variants of Concern, Variants of Interest and Variants Under Monitoring globally over the last several months. The World Health Organization (WHO) employs the Greek alphabet to characterize Variants of Concern (e.g., Alpha, Beta, Gamma, Delta) and Variants of Interest (e.g., Lambda, Mu). There is variation in variant dominance globally and the Delta Variant became dominant in the United States in summer 2021. The Delta Variant was characterized by WHO as a Variant of Interest on April 4, 2021 and was reclassified as a Variant of Concern on May 11, 2021. April Heyward has been examining COVID-19 and public policy interventions as part of her research, and employed machine learning and Twitter data to evaluate the public opinion of the Delta Variant and the Vaccine Booster. Public administration can employ computational methods such as machine learning to evaluate public opinion and policy interventions. This serves as a nod to computational social science.
Machine learning is a fascinating subdiscipline of artificial intelligence and its vastness facilities natural language processing to evaluate public policy and public opinions. Social media data is a treasure trove for data scientists, and coupled with machine learning, can extend the arsenal of public administration. Heyward started extracting Twitter data via RStudio (an Integrated Development Environment for the R Programming language) with the rtweet package developed by Michael Kearney to study the public opinions of COVID-19, including prevention and risk management measures, and public policy interventions in February 2021. Retweets are excluded from data extraction. Datasets are created with each extraction and are saved as .CSV files. The extraction of Delta Variant Twitter data began as the variant became dominant in the United States and the extraction of Vaccine Booster Twitter data began as formal deliberations of the availability of the booster shot commenced. The text of tweets is cleaned in RStudio by removing punctuations, http links, stopwords, digits, other symbols and capital letters which are transformed to lowercase letters for preparation of the emotions analysis. Lexicons are dictionaries that ascribe words to positive and negative sentiments and emotions. Employing multiple lexicons can yield varying results that can extend understanding of sentiments and emotions. The NRC lexicon, developed by Saif Mohammad and Peter Turney, and the Bing lexicon, developed by Bing Lui, were employed in the emotions analysis of the Delta Variant and Vaccine Booster tweets. There are numerous methods for evaluating tweets that are inclusive of, but not limited to, the analysis of the sources of tweets, positive and negative emotions found in tweets and words that contribute to positive and negative emotions.
The Delta Variant Twitter data employed in the emotions analysis was extracted on September 24, 2021 and yielded 17,918 tweets from the September 20-24, 2021 time period. A subsequent Delta Variant Twitter dataset was created retaining tweets resulting from Twitter for Android, Twitter for iPad, Twitter for iPhone, Twitter for Mac and Twitter Web App sources. This step reduced the number of tweets to 13,894. The Vaccine Booster Twitter data employed in the emotions analysis was extracted on the same day and yielded 17,615 tweets from the September 18-24, 2021 time period. A subsequent Vaccine Booster Twitter dataset was created, retaining tweets from the same Tweet sources and netted 12,424 tweets. See Figure 1 for the Delta Variant No Hashtag Tweets by Tweet Source Donut Chart. The Twitter Web App was the top source for Delta Variant tweets at 40.02% and then Twitter for iPhone (31.92%), Twitter for Android (23.76%), Twitter for iPad (4.18%) and Twitter for Mac (0.12%) respectively. See Figure 2 for Words Contributed to Positive and Negative Emotions Found in Delta Variant No Hashtag Tweets. Figure 2 depicts the emotions and words that contributed to the emotions found in Delta Variant tweets. The words found in Delta Variant tweets ascribed more to the positive emotion than the negative emotion when using the NRC lexicon. The word “vaccine” emerged as the most used and was ascribed to the positive emotion.
Figure 1
Figure 2
The vaccine booster is an intervention to continue combating COVID-19. See Figure 3 for Vaccine Booster No Hashtag Organic Tweets—Emotions Analysis of September 18-24, 2021 Tweets—NRC Lexicon. The positive emotion emerged as the leading emotion in the analysis as compared to other emotions. Employing the Bing lexicon to identify words contributing to positive and negative emotions found in Vaccine Booster tweets added more context to the results. See Figure 4 for Words Contributed to Positive and Negative Emotions Found in Vaccine Booster No Hashtag Tweets. The top ten words were identified in the Vaccine Booster tweets that contributed to positive and negative emotions. Machine learning is a method for public administration to adapt from the computer science discipline and to employ.
Figure 3
Figure 4
Author: April Heyward is an Author for PA TIMES, Public Sector Practitioner in South Carolina, 4th Year Doctor of Public Administration Student at Valdosta State University, and a R Programmer. She can be reached at [email protected] and followed on Twitter: https://twitter.com/heyward_april. All opinions and views are her own and does not reflect the views and opinions of her affiliations.
Follow Us!