Related Tweets Research Articles

Efficient Turkish Text Classification Approach for Crisis Management Systems

In this paper, an effective tweet classification system that fully supports the Turkish language has been developed. The proposed system can be used for mining (classifying) the recently published and publicly available tweets to find the crisis’s most related and useful tweets to gain situational awareness, which can help in taking the correct responses in order to prevent or at least decrease the effect of such situations. A deep study was carried out to improve and optimize the proposed system. In more detail, some intensive experiments were performed to investigate the performance of some well-known machine learning algorithms, i.e., K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Naive Bayes (NB) when used for text (tweets) classification. Then, the performances of the ensemble systems of the studied algorithms and the Random Forest (RF), AdaBoost Classifier (AdaBoost), GradientBoosting Classifier (GBC) ensemble systems have also been observed. As shown in the experimental evaluation and analysis, the proposed approach has stability, robustness, and can achieve quite good performance when processing the Turkish language. The performance of the proposed classifier was also compared with two state-of-the-art text classification approaches, i.e., "Empirical" and “Turkish Deep ".

Analyzing Impact of Social Media Sentiments on Financial Markets

This paper analyzes the impact of continuously changing sentiments on apparently unstable stock exchange. Right when a monetary supporter decides to buy or sell stock, his decision is very much dependent on to rise or fall in price of the stock. In this paper, we look at the possibility of using notion attitudes (good versus negative) and moreover sentiments (delight, feel sorry for, etc) isolated from finance related news or tweets to help predict stock worth turns of events. This examination uses a model-self-ruling approach to manage uncover the mysterious components of stock exchange data using distinctive significant learning techniques like Recurrent Neural Networks (RNN), Long-Short Term Memory (LSTM), and Gated Recurrent Unit (GRU).

Changes in Public Response Associated With Various COVID-19 Restrictions in Ontario, Canada: Observational Infoveillance Study Using Social Media Time Series Data.

BackgroundNews media coverage of antimask protests, COVID-19 conspiracies, and pandemic politicization has overemphasized extreme views but has done little to represent views of the general public. Investigating the public’s response to various pandemic restrictions can provide a more balanced assessment of current views, allowing policy makers to craft better public health messages in anticipation of poor reactions to controversial restrictions.ObjectiveUsing data from social media, this infoveillance study aims to understand the changes in public opinion associated with the implementation of COVID-19 restrictions (eg, business and school closures, regional lockdown differences, and additional public health restrictions, such as social distancing and masking).MethodsCOVID-19–related tweets in Ontario (n=1,150,362) were collected based on keywords between March 12 and October 31, 2020. Sentiment scores were calculated using the VADER (Valence Aware Dictionary and Sentiment Reasoner) algorithm for each tweet to represent its negative to positive emotion. Public health restrictions were identified using government and news media websites. Dynamic regression models with autoregressive integrated moving average errors were used to examine the association between public health restrictions and changes in public opinion over time (ie, collective attention, aggregate positive sentiment, and level of disagreement), controlling for the effects of confounders (ie, daily COVID-19 case counts, holidays, and COVID-19–related official updates).ResultsIn addition to expected direct effects (eg, business closures led to decreased positive sentiment and increased disagreements), the impact of restrictions on public opinion was contextually driven. For example, the negative sentiment associated with business closures was reduced with higher COVID-19 case counts. While school closures and other restrictions (eg, masking, social distancing, and travel restrictions) generated increased collective attention, they did not have an effect on aggregate sentiment or the level of disagreement (ie, sentiment polarization). Partial (ie, region-targeted) lockdowns were associated with better public response (ie, higher number of tweets with net positive sentiment and lower levels of disagreement) compared to province-wide lockdowns.ConclusionsOur study demonstrates the feasibility of a rapid and flexible method of evaluating the public response to pandemic restrictions using near real-time social media data. This information can help public health practitioners and policy makers anticipate public response to future pandemic restrictions and ensure adequate resources are dedicated to addressing increases in negative sentiment and levels of disagreement in the face of scientifically informed, but controversial, restrictions.

Leveraging Transfer Learning to Analyze Opinions, Attitudes, and Behavioral Intentions Toward COVID-19 Vaccines: Social Media Content and Temporal Analysis.

BackgroundThe COVID-19 vaccine is considered to be the most promising approach to alleviate the pandemic. However, in recent surveys, acceptance of the COVID-19 vaccine has been low. To design more effective outreach interventions, there is an urgent need to understand public perceptions of COVID-19 vaccines.ObjectiveOur objective was to analyze the potential of leveraging transfer learning to detect tweets containing opinions, attitudes, and behavioral intentions toward COVID-19 vaccines, and to explore temporal trends as well as automatically extract topics across a large number of tweets.MethodsWe developed machine learning and transfer learning models to classify tweets, followed by temporal analysis and topic modeling on a dataset of COVID-19 vaccine–related tweets posted from November 1, 2020 to January 31, 2021. We used the F1 values as the primary outcome to compare the performance of machine learning and transfer learning models. The statistical values and P values from the Augmented Dickey-Fuller test were used to assess whether users’ perceptions changed over time. The main topics in tweets were extracted by latent Dirichlet allocation analysis.ResultsWe collected 2,678,372 tweets related to COVID-19 vaccines from 841,978 unique users and annotated 5000 tweets. The F1 values of transfer learning models were 0.792 (95% CI 0.789-0.795), 0.578 (95% CI 0.572-0.584), and 0.614 (95% CI 0.606-0.622) for these three tasks, which significantly outperformed the machine learning models (logistic regression, random forest, and support vector machine). The prevalence of tweets containing attitudes and behavioral intentions varied significantly over time. Specifically, tweets containing positive behavioral intentions increased significantly in December 2020. In addition, we selected tweets in the following categories: positive attitudes, negative attitudes, positive behavioral intentions, and negative behavioral intentions. We then identified 10 main topics and relevant terms for each category.ConclusionsOverall, we provided a method to automatically analyze the public understanding of COVID-19 vaccines from real-time data in social media, which can be used to tailor educational programs and other interventions to effectively promote the public acceptance of COVID-19 vaccines.

Timing Tweets to Increase Effectiveness of Information Campaigns

Microblogging websites such as Twitter are increasingly being used by businesses/campaigners for timely dissemination of information to their followers. The diffusion of a tweet depends on several factors: the activity of the follower nodes, the responsiveness of follower nodes to tweets from the source node, the out-degree of the follower nodes, the content of recent related tweets seen by the follower node, etc. Using such factors, in this paper, we propose a framework to measure the effectiveness of an information campaign over Twitter. We consider a positive as well as a negative metric to measure the impact of a tweet: while retweets are used to measure the positive impact, the lack of a timely response from an active follower node is taken as a potential negative impact. We investigate the scheduling of tweets to increase the net positive impact while keeping the net negative impact below a desired level. We propose and study several scheduling algorithms by casting the problem in a Markov Decision Process (MDP) framework. In order to compare our algorithms, we estimate the model parameters from tweet data collected using the Twitter API from an arbitrarily selected node and its 6837 followers over several months. For this dataset, we find that if successive tweets in the campaign are novel, then substantial gains over user activity based scheduling can be obtained by scheduling tweets in time slots where the ratio of the expected positive and negative metrics is high. We call this the MaxRatio policy and we show that it is optimal under certain conditions. In cases where we are not certain about the response of users to successive related tweets, we identify another algorithm (which we call MaxReach) as a robust alternative.

You Are What You Tweet: Analyzing Twitter for Public Health

Analyzing user messages in social media can measure different population characteristics, including public health measures. For example, recent work has correlated Twitter messages with influenza rates in the United States; but this has largely been the extent of mining Twitter for public health. In this work, we consider a broader range of public health applications for Twitter. We apply the recently introduced Ailment Topic Aspect Model to over one and a half million health related tweets and discover mentions of over a dozen ailments, including allergies, obesity and insomnia. We introduce extensions to incorporate prior knowledge into this model and apply it to several tasks: tracking illnesses over times (syndromic surveillance), measuring behavioral risk factors, localizing illnesses by geographic region, and analyzing symptoms and medication usage. We show quantitative correlations with public health data and qualitative evaluations of model output. Our results suggest that Twitter has broad applicability for public health research.

A novel domain and event adaptive tweet augmentation approach for enhancing the classification of crisis related tweets

One of the purposes of detecting the crisis related tweets is the ability to single out the tweets that provide information about the helps needed and offered. Classification of such tweets is difficult because of the unavailability of sufficient annotated tweets in those categories. To facilitate such classifications, a domain and event adaptive augmentation approach is proposed. The main objective of the research is to enhance the classification of crisis related tweets that have less training samples. The proposed algorithms are designed to integrate the innate domain and event information during the selection of words for augmentation. Components such as CrisisLex lexicon, Word2Vec embeddings and WordNet are utilized for the proposed augmentation. Experimentation is carried out to substantiate the benefits of augmentation. Results indicate increased performance of the classifier when provided with the expanded dataset including the augmented and original tweets. To combat the problem of overfitting and class imbalance arising due to the lesser training samples, a novel tweets augmentation algorithm can be utilized. The advantage in the proposed algorithms is the ability to retain the structure and inherent nature of the tweets during the augmentation.

Wisdom of crowds detects COVID-19 severity ahead of officially available data

During the unfolding of a crisis, it is crucial to forecast its severity at an early stage , yet access to reliable data is often challenging early on. The wisdom of crowds has been effective at forecasting in similar scenarios. We investigated whether the initial regional social media reaction to the emerging COVID-19 pandemic in three critically affected countries has significant relations with their observed mortality a month later. We obtained COVID-19 related regionally geolocated tweets from Italian, Spanish, and United States regions. We quantified the predictive power of the wisdom of the crowds using correlations and regressions of geolocated Tweet Intensity (TI) during the initial social media attention peak versus the cumulative number of deaths a month ahead. We found that the intensity of initial COVID-19 related tweet attention at the beginning of the pandemic across Italian, Spanish, and United States regions is significantly related (p < 0.001) to the extent to which these regions had been affected by the pandemic a month later. This association is most striking in Italy as when at its peak of TI in late February 2020 only two of its regions had reported mortality. The collective wisdom of the crowds at early stages of the pandemic, when information on the number of infections was not broadly available, strikingly predicted the extent of mortality reflecting the regional severity of the pandemic almost a month later. Our findings could underpin the creation of real-time novelty detection systems aimed at early reporting of the severity of crises impacting a territory leading to early activation of control measures at a stage when available data is extremely limited.

Moral Expressions in 280 Characters or Less: An Analysis of Politician Tweets Following the 2016 Brexit Referendum Vote.

Ideas about morality are deeply entrenched into political opinions. This article examines the online communication of British parliamentarians from May 2017-December 2019, following the 2016 referendum that resulted in Britain's exit (Brexit) from the European Union. It aims to uncover how British parliamentarians use moral foundations to discuss the Brexit withdrawal agreement on Twitter, using Moral Foundations Theory as a classification basis for their tweets. It is found that the majority of Brexit related tweets contain elements of moral reasoning, especially relating to the foundations of Authority and Loyalty. There are common underlying foundations between parties, but parties express opposing viewpoints within a single foundation. The study provides useful insights into Twitter’s use as an arena for moral argumentation, as well as uncovers the politician’s uses of moral arguments during Brexit agreement negotiations on Twitter. It contributes to the limited body of work focusing on the moral arguments made by politicians through Twitter.

Language and Sentiment Regarding Telemedicine and COVID-19 on Twitter: Longitudinal Infodemiology Study.

BackgroundThe COVID-19 pandemic has necessitated a rapid shift in how individuals interact with and receive fundamental services, including health care. Although telemedicine is not a novel technology, previous studies have offered mixed opinions surrounding its utilization. However, there exists a dearth of research on how these opinions have evolved over the course of the current pandemic.ObjectiveThis study aims to evaluate how the language and sentiment surrounding telemedicine has evolved throughout the COVID-19 pandemic.MethodsTweets published between January 1, 2020, and April 24, 2021, containing at least one telemedicine-related and one COVID-19–related search term (“telemedicine-COVID”) were collected from the Twitter full archive search (N=351,718). A comparator sample containing only COVID-19 terms (“general-COVID”) was collected and sampled based on the daily distribution of telemedicine-COVID tweets. In addition to analyses of retweets and favorites, sentiment analysis was performed on both data sets in aggregate and within a subset of tweets receiving the top 100 most and least retweets.ResultsTelemedicine gained prominence during the early stages of the pandemic (ie, March through May 2020) before leveling off and reaching a steady state from June 2020 onward. Telemedicine-COVID tweets had a 21% lower average number of retweets than general-COVID tweets (incidence rate ratio 0.79, 95% CI 0.63-0.99; P=.04), but there was no difference in favorites. A majority of telemedicine-COVID tweets (180,295/351,718, 51.3%) were characterized as “positive,” compared to only 38.5% (135,434/351,401) of general-COVID tweets (P<.001). This trend was also true on a monthly level from March 2020 through April 2021. The most retweeted posts in both telemedicine-COVID and general-COVID data sets were authored by journalists and politicians. Whereas the majority of the most retweeted posts within the telemedicine-COVID data set were positive (55/101, 54.5%), a plurality of the most retweeted posts within the general-COVID data set were negative (44/89, 49.4%; P=.01).ConclusionsDuring the COVID-19 pandemic, opinions surrounding telemedicine evolved to become more positive, especially when compared to the larger pool of COVID-19–related tweets. Decision makers should capitalize on these shifting public opinions to invest in telemedicine infrastructure and ensure its accessibility and success in a postpandemic world.

Geospatial analysis of misinformation in COVID-19 related tweets

COVID-19 has emerged as a global pandemic caused by its highly transmissible nature during the incubation period. In the absence of vaccination, containment is seen as the best strategy to stop virus diffusion. However, public awareness has been adversely affected by discourses in social media that have downplayed the severity of the virus and disseminated false information. This article investigates COVID-19 related Twitter activity in May and June 2020 to examine the origin and nature of misinformation and its relationship with the COVID-19 incidence rate at the state and county level. A geodatabase of all geotagged COVID-19 related tweets was compiled. Multiscale Geographically Weighted Regression was employed to examine the association between social media activity and the spatial variability of disease incidence. Findings suggest that MGWR could explain 80% of the COVID-19 incidence rate variations indicating a strong spatial relationship between social media activity and spread of the Covid-19 virus. Discourse analysis was conducted on tweets to index tweets downplaying the pandemic or disseminating misinformation. Findings indicate that sites of Twitter misinformation showed more resistance to pandemic management measures in May and June 2020 later experienced a rise in the number of cases in July.

Studying Moral-based Differences in the Framing of Political Tweets

In this paper, we study the moral framing of political content on Twitter. Specifically, we examine differences in moral framing in two datasets: (i) tweets from US-based politicians annotated with political affiliation and (ii) COVID-19 related tweets in German from followers of the leaders of the five major Austrian political parties. Our research is based on recent work that introduces an unsupervised approach to extract framing bias and intensity in news using a dictionary of moral virtues and vices. In this paper, we use a more extensive dictionary and adapt it to German-language tweets. Overall, in both datasets, we observe a moral framing that is congruent with the public perception of the political parties. In the US dataset, democrats have a tendency to frame tweets in terms of care, while loyalty is a characteristic frame for republicans. In the Austrian dataset, we find that the followers of the governing conservative party emphasize care, which is a key message and moral frame in the party's COVID-19 campaign slogan. Our work complements existing studies on moral framing in social media. Also, our empirical findings provide novel insights into moral-based framing on COVID-19 in Austria.

COVID-19 outbreak: An ensemble pre-trained deep learning model for detecting informative tweets

On 11 March 2020, the (WHO) World Health Organization declared COVID-19 (CoronaVirus Disease 2019) as a pandemic. A further crisis has manifested mass fear and panic, driven by lack of information, or sometimes outright misinformation, alongside the coronavirus pandemic. Twitter is one of the prominent and trusted social media in this current outbreak. Over time, boundless COVID-19 headlines and vast awareness have been spreading, with tweets, updates, videos, and explosive posts. Few studies have been performed on the pandemic to detect and interrelate various disease types, including current coronavirus. However, it is pretty tricky to discriminate and detect a specific category. This work is motivated by the need to inform society about limiting irrelevant information and avoiding spreading negative emotions. In this context, the current work focuses on informative tweet detection in the pandemic to provide relevant information to the government, medical organizations, victims services, etc. This paper used a Majority Voting technique-based Ensemble Deep Learning (MVEDL) model. This MVEDL model is used to identify COVID-19 related (INFORMATIVE) tweets. The state-of-art deep learning models RoBERTa, BERTweet, and CT-BERT are used for best performance with the MVEDL model. The “COVID-19 English labeled tweets” dataset is used for training and testing the MVEDL model. The MVEDL model has shown 91.75 percent accuracy, 91.14 percent F1-score and outperforms the traditional machine learning and deep learning models. We also investigate how to use the MVEDL model for sentiment analysis on 226668 unlabeled COVID-19 tweets and their informative tweets. The application section discussed a comprehensive analysis of both actual and informative tweets. According to our knowledge, this is the first work on COVID-19 sentiment analysis using a deep learning ensemble model.

Understanding Behavioral Intentions Toward COVID-19 Vaccines: Theory-Based Content Analysis of Tweets.

BackgroundAcceptance rates of COVID-19 vaccines have still not reached the required threshold to achieve herd immunity. Understanding why some people are willing to be vaccinated and others are not is a critical step to develop efficient implementation strategies to promote COVID-19 vaccines.ObjectiveWe conducted a theory-based content analysis based on the capability, opportunity, motivation–behavior (COM-B) model to characterize the factors influencing behavioral intentions toward COVID-19 vaccines mentioned on the Twitter platform.MethodsWe collected tweets posted in English from November 1-22, 2020, using a combination of relevant keywords and hashtags. After excluding retweets, we randomly selected 5000 tweets for manual coding and content analysis. We performed a content analysis informed by the adapted COM-B model.ResultsOf the 5000 COVID-19 vaccine–related tweets that were coded, 4796 (95.9%) were posted by unique users. A total of 97 tweets carried positive behavioral intent, while 182 tweets contained negative behavioral intent. Of these, 28 tweets were mapped to capability factors, 155 tweets were related to motivation, 23 tweets were related to opportunities, and 74 tweets did not contain any useful information about the reasons for their behavioral intentions (κ=0.73). Some tweets mentioned two or more constructs at the same time. Tweets that were mapped to capability (P<.001), motivation (P<.001), and opportunity (P=.03) factors were more likely to indicate negative behavioral intentions.ConclusionsMost behavioral intentions regarding COVID-19 vaccines were related to the motivation construct. The themes identified in this study could be used to inform theory-based and evidence-based interventions to improve acceptance of COVID-19 vaccines.

Twitter Sentiment Analysis of Mobile Reviews using kernelized SVM

Sentiment analysis is a technique of analysing the opinions commented in social media on various topics. There are few ways the sentiment analysis can be done. Machine learning plays a crucial role for analysing the opinions and reviews. Mobile related tweets have been scraped from twitter. Noise on the tweets has been removed using pre-processing and feature vectors were created. Support Vector Machine has been used to classify the reviews as either positive or negative. 4 cross validation technique were used to bring out the better accuracy. 3 different sizes of dataset with iphone11 reviews have been used for training and testing with different kernels of SVM. RBF kernel is found to be working better for classifying the tweets but at the same time it has been found that the accuracy decreases when the data grow bigger.

Texas Public Agencies' Tweets and Public Engagement During the COVID-19 Pandemic: Natural Language Processing Approach.

BackgroundThe ongoing COVID-19 pandemic is characterized by different morbidity and mortality rates across different states, cities, rural areas, and diverse neighborhoods. The absence of a national strategy for battling the pandemic also leaves state and local governments responsible for creating their own response strategies and policies.ObjectiveThis study examines the content of COVID-19–related tweets posted by public health agencies in Texas and how content characteristics can predict the level of public engagement.MethodsAll COVID-19–related tweets (N=7269) posted by Texas public agencies during the first 6 months of 2020 were classified in terms of each tweet’s functions (whether the tweet provides information, promotes action, or builds community), the preventative measures mentioned, and the health beliefs discussed, by using natural language processing. Hierarchical linear regressions were conducted to explore how tweet content predicted public engagement.ResultsThe information function was the most prominent function, followed by the action or community functions. Beliefs regarding susceptibility, severity, and benefits were the most frequently covered health beliefs. Tweets that served the information or action functions were more likely to be retweeted, while tweets that served the action and community functions were more likely to be liked. Tweets that provided susceptibility information resulted in the most public engagement in terms of the number of retweets and likes.ConclusionsPublic health agencies should continue to use Twitter to disseminate information, promote action, and build communities. They need to improve their strategies for designing social media messages about the benefits of disease prevention behaviors and audiences’ self-efficacy.

You can simply rely on communities for a robust characterization of stances

We show that the structure of communities in social me- dia provides robust information for weakly supervised approaches to assign stances to tweets. Using as seed the SemEval 2016 Stance Detection Task annotated data, we retrieved a high number of topically related tweets. We then propagated information from the manually an- notated seed to the retrieved tweets and thus obtained a bigger training corpus. Classifiers trained with this bigger, weakly supervised dataset reach similar or better performance than those trained with the manually annotated seed. In addition, they are more robust with respect to common manual annotator errors or biases and they have arguably more coverage than smaller datasets. Weakly supervised approaches, most notably self- supervision, commonly suffer from error propagation. Interestingly, communities seem to provide a structure that constrains error propagation. In particular, weakly supervised classifiers that incorporate community struc- ture are more robust with respect to class imbalance. Additionally, this is a straightforward, transparent ap- proach, using standard tools and pipelines, cheaper and faster than methods like crowd sourcing for manual an- notations. Thus it facilitates adoption, interpretability and accountability.

Breaking community boundary: Comparing academic and social communication preferences regarding global pandemics.

The global spread of COVID-19 has caused pandemics to be widely discussed. This is evident in the large number of scientific articles and the amount of user-generated content on social media. This paper aims to compare academic communication and social communication about the pandemic from the perspective of communication preference differences. It aims to provide information for the ongoing research on global pandemics, thereby eliminating knowledge barriers and information inequalities between the academic and the social communities. First, we collected the full text and the metadata of pandemic-related articles and Twitter data mentioning the articles. Second, we extracted and analyzed the topics and sentiment tendencies of the articles and related tweets. Finally, we conducted pandemic-related differential analysis on the academic community and the social community. We mined the resulting data to generate pandemic communication preferences (e.g., information needs, attitude tendencies) of researchers and the public, respectively. The research results from 50,338 articles and 927,266 corresponding tweets mentioning the articles revealed communication differences about global pandemics between the academic and the social communities regarding the consistency of research recognition and the preferences for particular research topics. The analysis of large-scale pandemic-related tweets also confirmed the communication preference differences between the two communities.

USE OF CYBER HATE IN THE ELECTORAL CAMPAIGNS BY THE MAINSTREAM POLITICAL PARTIES OF PAKISTAN

Purpose of the study: The purpose of the study is to examine the use of cyber hate by the Pakistan’s mainstream political parties. The issue of poll rigging in Pakistan’s General Elections 2013 is examined through discourse analysis of the related tweets. The study also aims at comprehending the extent to which cyber ethics were violated during the digital electoral campaigns.&#x0D; Methodology: Discourse Analysis of the tweets generated from the official Twitter handles of PTI and PMLN leaders was conducted to examine the use of cyber hate by the Pakistan’s mainstream political parties. Violation of cyber ethics was explored through the qualitative interviews of 8 purposively selected social media managers of PMLN, PPP, and PTI.&#x0D; Main Findings: The findings indicated that party leadership/politicians used the elements of cyber hate which included abusive language, provocation, and character assassination against their opponents during the digital electoral campaign in general and regarding the poll rigging issue of Pakistan’s General Elections 2013 in specific. Resultantly the tweets using strong adjectives and metaphors on the political opponents were more frequently re-tweeted and attracted more favorites.&#x0D; Applications of this study: The study can be helpful in various cross-disciplinary areas that focus on the examination of the usage and impact of social media and cyberspace as a medium for hate speech dissemination. The study can significantly contribute to areas related to cyber ethics, digital electoral campaigning, freedom of expression, and political opinion building.&#x0D; Novelty/Originality of this study: The study’s originality lies in its attempt to unfold the foundations of digital electoral campaigning in Pakistan and how cyberhate was used as a pivotal tool for advancing the political narratives in a fragile democratic society.

Examining Tweet Content and Engagement of Canadian Public Health Agencies and Decision Makers During COVID-19: Mixed Methods Analysis.

BackgroundEffective communication during a health crisis can ease public concerns and promote the adoption of important risk-mitigating behaviors. Public health agencies and leaders have served as the primary communicators of information related to COVID-19, and a key part of their public outreach has taken place on social media platforms.ObjectiveThis study examined the content and engagement of COVID-19 tweets authored by Canadian public health agencies and decision makers. We propose ways for public health accounts to adjust their tweeting practices during public health crises to improve risk communication and maximize engagement.MethodsWe retrieved data from tweets by Canadian public health agencies and decision makers from January 1, 2020, to June 30, 2020. The Twitter accounts were categorized as belonging to either a public health agency, regional or local health department, provincial health authority, medical health officer, or minister of health. We analyzed trends in COVID-19 tweet engagement and conducted a content analysis on a stratified random sample of 485 tweets to examine the message functions and risk communication strategies used by each account type.ResultsWe analyzed 32,737 tweets authored by 118 Canadian public health Twitter accounts, of which 6982 tweets were related to COVID-19. Medical health officers authored the largest percentage of COVID-19–related tweets (n=1337, 35%) relative to their total number of tweets and averaged the highest number of retweets per COVID-19 tweet (112 retweets per tweet). Public health agencies had the highest frequency of daily tweets about COVID-19 throughout the study period. Compared to tweets containing media and user mentions, hashtags and URLs were used in tweets more frequently by all account types, appearing in 69% (n=4798 tweets) and 68% (n=4781 tweets) of COVID-19–related tweets, respectively. Tweets containing hashtags also received the highest average retweets (47 retweets per tweet). Our content analysis revealed that of the three tweet message functions analyzed (information, action, community), tweets providing information were the most commonly used across most account types, constituting 39% (n=181) of all tweets; however, tweets promoting actions from users received higher than average retweets (55 retweets per tweet). When examining tweets that received one or more retweet (n=359), the difference between mean retweets across the message functions was statistically significant (P<.001). The risk communication strategies that we examined were not widely used by any account type, appearing in only 262 out of 485 tweets. However, when these strategies were used, these tweets received more retweets compared to tweets that did not use any risk communication strategies (P<.001) (61 retweets versus 13 retweets on average).ConclusionsPublic health agencies and decision makers should examine what messaging best meets the needs of their Twitter audiences to maximize sharing of their communications. Public health accounts that do not currently employ risk communication strategies in their tweets may be missing an important opportunity to engage with users about the mitigation of health risks related to COVID-19.