The missing link between Network Science and the Social Media Monitoring industry Pablo Arag´ on Barcelona Media, Spain State of the art of the Social Media Monitoring Industry In the last decade, social media have exploded as a key channel for the communication among people around the world. The communication/marketing industry has reacted to this trend by implementing thousands of Social Media Monitoring (SMM) tools to measure online conversations. The comparative of 245 SMM major tools conducted in [1] reveals interesting insights into the current state-of-the-art features for data analysis and visualization with a main focus on sentiment analysis, influencer profiling, viral content tracking and trend/topic analysis. The report exposes that most tools base their algorithmics on text analytics through Natural Language Processing methods or trivial approaches that count and rank occurrences (e.g., most retweeted users, trending topics, etc.). Consequently, data visualization is mainly covered by dashboards composed of lists of ranked items, line/pie/bar charts and word/tag clouds. At this point, it might be interesting to reflect on how these techniques are typical features of media analysis but ignore the social aspect of social media. If social media allow people to interact in virtual networks, should Network Science not play a primary role in the SMM industry? Over the last decades, the academia has deepened in the development and refinement of network models and algorithms for a large range of goals such as community detection, node centrality or link analysis. Many of these techniques can be found in numerous open-source software packages (e.g., NWB, SNAP, Pajek, Gephi, iGraph) and tools for social scientists (e.g., the platforms developed by Truthy from Indiana University and the Digital Methods Initiative from the University of Amsterdam). However, most industrial SMM platforms hardly integrate such techniques1 . Is this missing link between Network Science and the SMM industry not a relevant challenge to be addressed? Assessing KALIUM, a candidate solution The most simple solution to this problem is to integrate Social Network Analysis (SNA) techniques into SMM processes. This philosophy has inspired KALIUM, a SMM system that combines (a) dashboards similar to the ones exposed in [1] and (b) SNA techniques. Figure 1 illustrates an example for the case-study of the Mobile World Congress 2015 on Twitter. On the left side, a dashboard ranks ZTE Corporation as the 4th most retweeted user, as other SMM tools reported2 . On the right side, the network analysis on a specific day reveals that, in spite of its high-in-degree, ZTE Corporation was mostly poked by a huge cluster of users who did not belong to the giant component of the graph (led by the organizers of the event, mobile providers and mass-media). The reason behind this unexpected scenario is that the ZTE retweet-graph community was formed by bots, a common practice hardly detected by most SMM industrial tools. This basic example highlights the superficial level of the analysis performed by the SMM industry ignoring Network Science techniques. Despite the doubtless value of content analysis in SMM processes, the strengh of the link between Network Science and the SMM industry will be crucial for the social comprehension of social media. Figure 1. Social Media Monitoring of the Mobile World Congress 2015 on Twitter with KALIUM system. References [1] Milic Luisa. Social media monitoring tools and services report public excerpts 2014. Ideya Market Report, 5, 2014. 1 There are some minor and interesting exceptions like Maven7 (http://maven7.com/) or Lynguo (http://www.iic.uam.es/). press release: http://www.tynmagazine.com/tras-la-tercera-jornada-decrece-el-interes-en-twitter-sobre-el-mwc2015/ 2 Spanish 1
© Copyright 2024