Measuring User Influence in Twitter: The Million Follower Fallacy

“Furthermore , because of the tight connection between users as suggested in the triadic closure (Granovetter 1973), retweeting in a social network can serve as a powerful tool to reinforce a message—for instance, the probability of adopting an innovation increases when not one but a group of users…”

“Google similarly normalizes the data when analyzing their search trends (Ginsberg et al. 2009).”

“The notion of influence plays a vital role in how businesses operate and how a society functions—for instance, see observations on how fashion spreads (Gladwell 2002) and how people vote (Berry and Keller 2003).”

“Instead, it posits that the key factors determining influence are (i) the interpersonal relationship among ordinary users and (ii) the readiness of a society to adopt an innovation (Watts and Dodds 2007; Domingos and Richardson 2001).”

“Timing is related to making information quickly accessible - the earlier the information reaches someone the more value this person can get from it.”

“We used indegree centrality as a measure of influence in the network because it is based on the number of retweets, replies and mentions one receives in the conversation (Cha et al., 2010).”

“The Twitterderived features are widely-used to measure activity, status [16], and influence [5], and they are three: 1) the logarithm of the total number of tweets (activity); 2) the logarithm of the total number of followers divided by the number of followees (status); and 3) the sum of the average number of retweets…”

“Cha et al. [4] studied influence on Twitter.”

“2 Therefore, using solely the number of followers to quantify impact may lead to inaccurate outcomes (Cha et al., 2010).”

“…(Shaw et al., 2013); second , there have been studies trying to profile overall platform characteristics, often working with complete data-sets (e.g. Cha et al., 2010; Prieur et al., 2008) or representative samples (e.g. Gerlitz and Rieder, 2013); third , a considerable number of researchers,…”

“Early studies regarding influence on Twitter suggested that the most active and the most visible users may indicate influence on Twitter [73].”

“The Twitter dataset is extracted by unbiased sampling [Cha et al. 2010].”

“This is consistent with findings of other studies, as for example Cha et al. (2010) found that a high percentage of re-tweeting of tweets is an indicator of the level of influence of the users (who posted the messages) on their followers.”

“Marketing in particular found a perfect fit since now businesses are able to analyse a large volume of public data and identify trends [55], influ-ential profiles [18], experts [33] or to provide personalized advertisements and documents [20].”

“…the organisation on the platform, the number of followers and shares of content/calls made by the organisation and the popularity and stature of the crowds they engage (e.g. attention from celebrities, Nobel laureates), alongside the quality of the content posted (Cha et al. 2010; Taeihagh 2017a).”

“Generic approximation algorithms tend to provide small resolving sets but do not scale well enough with network size to be useful on larger networks which routinely include more than 10(6) total nodes [14, 83, 108].”

“The purpose of this two-phase design was to understand if users' responses would differ depending on whether they are followed, given prior evidences that some Twitter users would follow back their followers simply for etiquette (Cha et al. 2010).”

“From 2008, it is also a source of study by academic researchers [6] and it has been applied in a number of social fields such as elections [7] [8] [9], social movements [10], predictions [11] [12], user’s influence [13], behavior [14] [15], and message propagation analytics [16].”

“Even if in some online social services, such as Twitter, user behavior may appear more interest-driven than purely social (Kwak et al. 2010b; Cha et al. 2010), connections between online users represent rich source of information to understand these systems.”

“Sharing information (e.g., retweeting) is a key feature provided by social media sites such as Twitter (Cha et al., 2010; Suh et al., 2010).”

“Previous studies indicate that the quantity and speed of retweeting are a good measure for message virality and a screen-name’s influence (Cha et al., 2010; Hoang & Lim, 2012).”

“The result confirms several existing studies suggesting that high social influence requires long-term, consistent effort [44, 45].”

“widespread use of social media sites, like Twitter and Facebook, and the ever growing number of users, there has been much effort on understanding and modeling information propagation on social media [1, 2, 6, 14, 17, 25, 27, 29, 30]. Most of the work assumes that information is propagated by a small number of influen…”

“This echoes work by Cha et al. (2010) who provide evidence that having a large number of Twitter followers is not that predictive of influence on Twitter.”

“MO [21] Human Social 217 2,672 TC [2] Infrastructure 1,226 2,615 OF [42] Infrastructure 2,939 30,501 AD [38] Social 6,541 51,127 AM [35] E-commerce 403,394 3,387,388 AR [40] E-commerce 3,376,972 5,838,041 BA [41] Hyperlink 2,141,300 17,794,839 TW [9] Social 52,579,682 1,963,263,821”

“, to trigger collective attention (9, 10), gain status (11, 12), and monetize public attention (13)] or with malicious intent [e.”

“These two competing aspects have led to increased focus on Twitter influence: how much one user can affect others [2], [8], [9], and how different user networks are connected to spread information [5], [35].”

“For Twitter data, various influence measures include number of followers, mentions, retweets [4], Pagerank of follower network [20], size of the information cascades [2].”

“Table I lists the graphs used from the University of Florida Sparse Matrix Collection [3]– [5], [13], the 10th DIMACS Implementation Challenge web-site [1], the Stanford Network Analysis Platform (SNAP) website [23], [30], [35], the Koblenz Network Collection [22], and Cha et al. [9].”

“Many studies have discussed influencers on social media in general [1-5] and in the political context specifically [5].”

“The same has happened with Twitter, where the first massive studies (Cha et al. 2010; Kwak et al. 2010) have shown its nature not as a social network but as an information one.”

“While prior research focused on Twitter users who were highly influential in spreading information [8], [9], we focus on users who had the strongest possibility to block or disrupt the flow of information.”

“An influential user is the onewith a huge degree and many mentions and shares (retweets) [13].”

“Cha et al. (2010) stated that reciprocity (following-back action) often occurs not because of the social influence of the Twitter user but because of etiquette.”

“Previous research has explored the differences between tweets, retweets and replies on Twitter [31, 32].”

“The ability to communicate interesting political issues yields the opportunity to users to acquire more visibility and influence [2, 9, 43], although Twitter political discussion is plagued by a number of issues related to manipulation and abuse [21,22,45].”

“…Vliegenthart (2017), define influentials as “users with above average ability to stimulate retweets to their own messages” (p. 503), consistent with conceptualization of influencers based on impact on content distribution (Cha, Haddadi, Benevenuto, & Gummadi, 2010; Kwak, Lee, Park, & Moon, 2010).”

“[11] investigate the dynamics of user influence across topics and time based on an in-depth comparison of three measures of influence: in-degree, retweets, and mentions.”

“comparison between in-degree, re-tweets and mentions as influence indicators([14]) concluded that the first is more related to user popularity.”

“Users generate a constant stream of online messages through social media to share and discuss their activities, status, opinions, ideas and interesting news stories; social media might be an effective means to examine trends and popularity in topics ranging from economic, social, environmental to political issues [1],…”

“…information .ltering, retrieval models General Terms Algorithms, Design, Experimentation Keywords Information Networks, In.uence, Social Networks Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that…”

“[13] have examined three types of influential users, performed in propagating popular news topics.”

“…is based on previous work in network science to characterize online communities, as well as insights about the long-tailed distribution of social media followership in general (Cha et al. 2010), and it is a novel application of established statistical procedures in this context (Fog 2007).”

“A user’s network activity in social media means how each user influences the other users based on the number of Twitter users’ mentions and retweets [30].”

“is, instead, key. With regards to [17] and [47] the behavior of the users (either retweets or mentions) is as important as the social graph or, as is the case of [17], it simply replaces it. Finally, [8] just made a point against the use of simplistic graph measures (indegree) as proxies for influence but did not make…”

“When an edge ( i, j ) is inserted to and/or removed from a network at time t , only the PPR scores of the vertices “around” vertices i and j are updated.”

“The recall of the FansRank and RetweetInfluence show an upward trend with the increasing of k .”

“The highlighted route is the shortest path from i to j . c-d , Giant component G ( q ) of Twitter ( N = 469 , 014) and Mobile phone network in Mexico ( N = 1 .”

“” Clarke (2015) and Manor (2018b) similarly suggest that diplomacy should not be considered in binary terms—traditional or digital—as all diplomats are expected to incorporate digital tools in their traditional daily work.”

“In addition, it may contribute to a great increase or decrease in marketing, especially economic loss (Brock and Hommes, 1998; Duman and Huosheng, 2001; Cha et al., 2010).”

“Twitter-specific studies have operationalized influence by calculating how often a user is mentioned or retweeted (Cha & Gummadi, 2010) or by considering expert content in the messages and social embeddedness in a local network (Dubois & Gaffney, 2014).”

“3380052 to promote brand awareness and advertise products to social networks of influencers [13, 17, 22, 31], who have ‘influence’ over a large number of followers [3, 4], since customers are often more likely to trust influencers’ recommendations than brands’ advertisements [5, 26, 31].”

“performed a study on how influential people are related to the number of retweeted messages [4,5].”

“Lastly, personal influence has been widely studied for online social networks [42], [43].”

“Figure 3 shows the CPU time for all methods across tweet files of increasing sizes, which are drawn from large tweet dataset in [6].”

“Based on several studies, retweets are considered to be one of the best indicators of user popularity from the quantitative perspective [14,23,28].”

“Twitter has emerged as a critical factor in information dissemination, marketing [19], and influence discovery [9].”

“many following relationships on Twitter do not lead to retweets [5], ProfileRank is able to identify only those relationships that induce information diffusion on the network.”

“These studies have resulted in important findings about the effect of misinformation spread [33], influence campaigns [2, 7], and social bots [38] within specific newsworthy events.”

“By analyzing 1.7B tweets, obtained from [51], they find that: 1) users that spread rumors and non-rumors have similar registration age and number of followers; 2) rumors have a clearly different writing style; 3) sentiment in news depends on the topic and not on the credibility of the post; and 4)…”

“Social media have been widely portrayed as enablers of democracy [50,48,15,12,47].”

“…sobre los demás en Twitter no es algo que se consiga de forma puntual, requiere de un esfuerzo continuado y una alta implicación personal (Cha & Gummadi, 2010) y aunque las personas que son populares e influyentes tienen más probabilidad de influir en el futuro, no está garantizado que…”

“Cha et al. (2010) suggested 1128 J. Y. M. NIP AND K.-W. FU that follower size was a measure of user popularity, retweet count a measure of the tweet’s content value, and mention count an indication of the name value of the user.”

“We use three publicy available datasets for our experiments: a webgraph called Webbase(3) [5], which consists of 1,019,903,190 links between 115,657,290 webpages, a snapshot of Twitter’s social network(4) [10], which contains 1,963,263,821 follower links between 51,217,936 users and a dataset of 717,872,016 ratings th…”