Validation of Twitter opinion trends with national polling aggregates: Hillary Clinton vs Donald Trump.
ABSTRACT: Measuring and forecasting opinion trends from real-time social media is a long-standing goal of big-data analytics. Despite the large amount of work addressing this question, there has been no clear validation of online social media opinion trend with traditional surveys. Here we develop a method to infer the opinion of Twitter users by using a combination of statistical physics of complex networks and machine learning based on hashtags co-occurrence to build an in-domain training set of the order of a million tweets. We validate our method in the context of 2016 US Presidential Election by comparing the Twitter opinion trend with the New York Times National Polling Average, representing an aggregate of hundreds of independent traditional polls. The Twitter opinion trend follows the aggregated NYT polls with remarkable accuracy. We investigate the dynamics of the social network formed by the interactions among millions of Twitter supporters and infer the support of each user to the presidential candidates. Our analytics unleash the power of Twitter to uncover social trends from elections, brands to political movements, and at a fraction of the cost of traditional surveys.
Project description:Social media has become an emerging alternative to opinion polls for public opinion collection, while it is still posing many challenges as a passive data source, such as structurelessness, quantifiability, and representativeness. Social media data with geotags provide new opportunities to unveil the geographic locations of users expressing their opinions. This paper aims to answer two questions: 1) whether quantifiable measurement of public opinion can be obtained from social media and 2) whether it can produce better or complementary measures compared to opinion polls. This research proposes a novel approach to measure the relative opinion of Twitter users towards public issues in order to accommodate more complex opinion structures and take advantage of the geography pertaining to the public issues. To ensure that this new measure is technically feasible, a modeling framework is developed including building a training dataset by adopting a state-of-the-art approach and devising a new deep learning method called Opinion-Oriented Word Embedding. With a case study of tweets on the 2016 U.S. presidential election, we demonstrate the predictive superiority of our relative opinion approach and we show how it can aid visual analytics and support opinion predictions. Although the relative opinion measure is proved to be more robust than polling, our study also suggests that the former can advantageously complement the latter in opinion prediction.
Project description:Previous research finds that voting is a socially stressful activity associated with increases in cortisol levels. Here we extend this research by investigating whether different voting modalities have differential effects on the stress response to voting. Results from a field experiment conducted during the 2012 presidential elections strongly suggest that traditional "at the polls" voting is more stressful, as measured by increases in cortisol levels, than voting at home by mail-in ballot or engaging in comparable non-political social activities. These findings imply that increased low-stress voting options such as mail-in ballots may increase political participation among individuals who are sensitive to social stressors.
Project description:Multiple countries have recently experienced extreme political polarization, which, in some cases, led to escalation of hate crime, violence and political instability. Besides the much discussed presidential elections in the USA and France, Britain's Brexit vote and Turkish constitutional referendum showed signs of extreme polarization. Among the countries affected, Ukraine faced some of the gravest consequences. In an attempt to understand the mechanisms of these phenomena, we here combine social media analysis with agent-based modelling of opinion dynamics, targeting Ukraine's crisis of 2014. We use Twitter data to quantify changes in the opinion divide and parametrize an extended bounded confidence XY model, which provides a spatio-temporal description of the polarization dynamics. We demonstrate that the level of emotional intensity is a major driving force for polarization that can lead to a spontaneous onset of collective behaviour at a certain degree of homophily and conformity. We find that the critical level of emotional intensity corresponds to a polarization transition, marked by a sudden increase in the degree of involvement and in the opinion bimodality.
Project description:BACKGROUND:Digital spaces, and in particular social networking sites, are becoming increasingly present and influential in the functioning of our democracies. In this paper, we propose an integrated methodology for the data collection, the reconstruction, the analysis and the visualization of the development of a country's political landscape from Twitter data. METHOD:The proposed method relies solely on the interactions between Twitter accounts and is independent of the characteristics of the shared contents such as the language of the tweets. We validate our methodology on a case study on the 2017 French presidential election (60 million Twitter exchanges between more than 2.4 million users) via two independent methods: the comparison between our automated political categorization and a human categorization based on the evaluation of a sample of 5000 profiles descriptions; the correspondence between the reconfigurations detected in the reconstructed political landscape and key political events reported in the media. This latter validation demonstrated the ability of our approach to accurately reflect the reconfigurations at play in the off-line political scene. RESULTS:We built on this reconstruction to give insights into the opinion dynamics and the reconfigurations of political communities at play during a presidential election. First, we propose a quantitative description and analysis of the political engagement of members of political communities. Second, we analyze the impact of political communities on information diffusion and in particular on their role in the fake news phenomena. We measure a differential echo chamber effect on the different types of political news (fake news, debunks, standard news) caused by the community structure and emphasize the importance of addressing the meso-structures of political networks in understanding the fake news phenomena. CONCLUSIONS:Giving access to an intermediate level, between sociological surveys in the field and large statistical studies (such as those conducted by national or international organizations) we demonstrate that social networks data make it possible to qualify and quantify the activity of political communities in a multi-polar political environment; as well as their temporal evolution and reconfiguration, their structure, their alliance strategies and their semantic particularities during a presidential campaign through the analysis of their digital traces. We conclude this paper with a comment on the political and ethical implications of the use of social networks data in politics. We stress the importance of developing social macroscopes that will enable citizens to better understand how they collectively make society and propose as example the "Politoscope", a macroscope that delivers some of our results in an interactive way.
Project description:BACKGROUND:Tobacco-related content on social media is generated and propagated by opinion leaders on the Web who disseminate messages to others in their network, including followers, who then continue to spread the information. Opinion leaders can exert powerful influences on their followers' knowledge, attitudes, and behaviors; yet, little is known about the demographic characteristics and tobacco use behavior of tobacco opinion leaders on the Web and their followers, compared with general Twitter users. OBJECTIVE:In this study, we hypothesized that opinion leaders use more tobacco products and have higher nicotine dependence than the other 2 groups (eg, followers and general Twitter users) and that followers-those who spread messages by opinion leaders-would more likely be in demographic groups that are vulnerable to tobacco marketing influence (eg, young adults and lower educational attainment). METHODS:We constructed the social networks of people who tweet about tobacco and categorized them using a combination of social network and Twitter metrics. To understand the characteristics of tobacco opinion leaders and their followers, we conducted a survey of tobacco opinion leaders, their followers, and general Twitter users. The sample included 347 opinion leaders, 567 followers, and 519 general users. The opinion leaders had a median of 1000 followers, whereas followers and general users had fewer than 600 followers. RESULTS:Opinion leaders were more likely than their followers to report past month use of tobacco products; followers, in turn, were more likely to report past month use of these products than general Twitter users. The followers appeared to be an especially vulnerable group; they tended to be younger (mean age 22.4 years) and have lower education compared with the opinion leaders and general users. CONCLUSIONS:Followers of Twitter tobacco opinion leaders are a vulnerable group that might benefit from antitobacco education to counter the protobacco communications they see on social media.
Project description:This paper investigates the relationship between candidates' online popularity and election results, as a step towards creating a model to forecast the results of Taiwanese elections even in the absence of reliable opinion polls on a district-by-district level. 253 of 354 legislative candidates of single-member districts in Taiwan's 2016 general election had active public Facebook pages during the election period. Hypothesizing that the relative popularity of candidates' Facebook posts will be positively related to their election results, I calculated each candidate's Like Ratio (i.e. proportions of all likes on Facebook posts obtained by candidates in their district). In order to have a measure of online interest without the influence of subjective positivity, I similarly calculated the proportion of daily average page views for each candidate's Wikipedia page. I ran a regression analysis, incorporating data on results of previous elections and available opinion poll data. I found the models could describe the result of the election well and reject the null hypothesis. My models successfully predicted 80% of winners in single-member districts and were effective in districts without local opinion polls with a predictive power approaching that of traditional opinion polls. The models also showed good accuracy when run on data for the 2014 Taiwanese municipal mayors election.
Project description:Despite recent and growing interest in using Twitter to examine human behavior and attitudes, there is still significant room for growth regarding the ability to leverage Twitter data for social science research. In particular, gleaning demographic information about Twitter users-a key component of much social science research-remains a challenge. This article develops an accurate and reliable data processing approach for social science researchers interested in using Twitter data to examine behaviors and attitudes, as well as the demographic characteristics of the populations expressing or engaging in them. Using information gathered from Twitter users who state an intention to not vote in the 2012 presidential election, we describe and evaluate a method for processing data to retrieve demographic information reported by users that is not encoded as text (e.g., details of images) and evaluate the reliability of these techniques. We end by assessing the challenges of this data collection strategy and discussing how large-scale social media data may benefit demographic researchers.
Project description:<i>Background and objectives:</i> Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has spread to more than 200 countries. In light of this situation, the Japanese Government declared a state of emergency in seven regions of Japan on 7 April 2020 under the provisions of the law. The medical care delivery system has been under pressure. Although various surgical societies have published guidelines on which to base their surgical decisions, it is not clear how general anesthesia has been performed and will be performed in Japan. <i>Materials and Methods:</i> One of the services provided by the social network service Twitter is a voting function-Twitter Polls-through which anonymous surveys were conducted. We analyzed the results of a series of surveys 17 times over 22 weeks on Twitter on the status of operating restrictions using quadratic programming to solve the mathematical optimizing problem, and public data provided by the Japanese Government were used to estimate the current changes in the number of general anesthesia performed in Japan. <i>Results:</i> The minimum number of general anesthesia cases per week was estimated at 67.1% compared to 2015 on 27 April 2020. The timeseries trend was compatible with the results reported by the Japanese Society of Anesthesiologists (correlation coefficient <i>r</i> = 0.69, <i>p <</i> 0.001<i>)</i>. <i>Conclusions:</i> The number of general anesthesia was reduced up to two-thirds during the pandemic of COVID-19 in Japan and was successfully quantitatively estimated using a quick questionnaire on Twitter.
Project description:The dynamics and influence of fake news on Twitter during the 2016 US presidential election remains to be clarified. Here, we use a dataset of 171 million tweets in the five months preceding the election day to identify 30 million tweets, from 2.2 million users, which contain a link to news outlets. Based on a classification of news outlets curated by www.opensources.co , we find that 25% of these tweets spread either fake or extremely biased news. We characterize the networks of information flow to find the most influential spreaders of fake and traditional news and use causal modeling to uncover how fake news influenced the presidential election. We find that, while top influencers spreading traditional center and left leaning news largely influence the activity of Clinton supporters, this causality is reversed for the fake news: the activity of Trump supporters influences the dynamics of the top fake news spreaders.
Project description:Recently we have witnessed a number of rapid shifts toward populism in the rhetoric and policies of major political parties, as exemplified in the 2016 Brexit Referendum, 2016 US Election, and 2017 UK General Election. Our perspective here is to focus on understanding the underlying societal processes behind these recent political shifts. We use novel methods to study social dynamics behind the 2016 Presidential election. This is done by using network science methods to identify key groups associated with the US right-wing during the election. We investigate how the groups grew on Twitter, and how their associated accounts changed their following behaviour over time. We find a new external faction of Trump supporters took a strong influence over the traditional Republican Party (GOP) base during the election campaign. The new group dominated the GOP group in terms of new members and endorsement via Twitter follows. Growth of new accounts for the GOP party all but collapsed during the campaign. While the Alt-right group was growing exponentially, it has remained relatively isolated. Counter to the mainstream view, we detected an unexpectedly low number of automated 'bot' accounts and accounts associated with foreign intervention in the Trump-supporting group. Our work demonstrates a powerful method for tracking the evolution of societal groups and reveals complex social processes behind political changes.