Social Media Mining
Social Media Mining is the process of extracting and analyzing vast amounts of data generated by user interactions on social media platforms such as Facebook, Twitter (now X), Instagram, and LinkedIn. This discipline focuses on understanding how individuals—referred to as social atoms—interact within virtual communities, known as social molecules. The enormous volume of data produced daily can be categorized as big data, presenting both opportunities and challenges for analysis due to its structured and unstructured nature.
The insights gained from Social Media Mining have diverse applications, including market research, social behavior analysis, and trend forecasting. Companies leverage this data to better understand consumer preferences and tailor their marketing strategies accordingly. Additionally, political analysts and government officials utilize mined data to gauge public sentiment and priorities. As technology advances, particularly with tools like deep learning and natural language processing, the accuracy and efficiency of social media mining continue to improve, allowing researchers and businesses to derive meaningful patterns from complex datasets in real-time. This growing field holds significant potential for understanding human behavior and enhancing decision-making processes across various sectors.
Subject Terms
Social Media Mining
Social media is made up of the numerous websites, online platforms, and applications where people share information while building virtual communities. Social media mining is the process of gathering and studying the immense amount of data collected from social media platforms such as Facebook, X (formerly known as Twitter), Instagram, Tumblr, and LinkedIn. Social media platforms generate a massive amount of data through the interactions of individuals with each other and their surroundings. In the context of social media mining, individuals are known as social atoms, and the communities they form are known as social molecules. The data collected from the interaction of social atoms and social molecules can be extracted and analyzed—a process known as mining—in order to realize meaningful patterns that can be examined and used.
Through mining social media, social scientists are able to study how individuals interact with each other and their environments. They are also able to see how virtual communities are formed and maintained. Data mining of social media is also used for marketing and other commercial purposes.
Background on Social Media
Since social media’s advent in the early 2000s, people have been using it to upload videos, images, news stories, personal information, and other content in vast amounts to webpages and applications for the purpose of sharing this content and thus interacting with others around the world. This information can be accessed immediately via computers or mobile devices. Viewers of the content can usually interact with the original poster by giving feedback through comments or by using onscreen features, such as Facebook’s "like" button, to signal their approval or share the content with others.
According to the website eBizMBA, which provides information on e-business, the most popular social networking platforms in May 2016 were Facebook, YouTube, Twitter, LinkedIn, Pinterest, Google Plus, Tumblr, Instagram, Reddit, VK, Flickr, and Vine. The sites’ popularity rankings were based on the estimated number of unique visitors each site receives on a monthly basis. Of these, Facebook was the most-used social media network, and the most recognized worldwide.
By 2024, some platforms remained popular, while others had fallen out of favor. According to the Pew Research Center, YouTube ranked the highest with 83 percent of adults ever having visited the site, while 68 percent reported using Facebook, 47 percent reported using Instagram, and 33 percent reported using TikTok. Despite the smaller percentage of adults who had visited TikTok, though, Pew reported that the platform had experienced a 21 percent growth since 2021.
Because social media has a global audience, the information distributed on these networks is extensive, with billions of pieces of data. This data is disorganized and sporadic. Special techniques and a multidisciplinary approach are needed in order to gather, analyze, and present the information in a useful way. The data that is mined can then be used for marketing strategies, productivity, increased revenue, and understanding human social behavior.
Impact of Social Media Mining
The data that is mined from social media is so vast and from so many sources that it is considered big data, which refers to data that exceeds petabytes (one million gigabytes) in size. Big data may be structured, meaning that it is identifiable and retrievable, or unstructured, meaning it cannot be easily retrieved or analyzed.
According to Forbes in 2018, roughly 2.5 quintillion bytes of data were created on an average day and the pace has continued to increase. Such a large amount of data cannot be analyzed by traditional data mining means. Database software on regular computers would take months to be able to make sense of any of it. That is why computer scientists must use supercomputers and new techniques of data mining to analyze and interpret much of the social media data that is generated.
Social media data differs from other data in several ways. For example, it relies on social relations and actions like friending and following, where users post content where others may comment on it, "like" it by clicking on a positive icon to highlight their approval, or repost it through sharing on their own social media pages. Therefore, making sense of the data requires a multidisciplinary approach to retrieving and interpreting it, one that combines the established technical algorithms and methods of data mining from computer science and mathematics with social network analysis, statistics, ethnography, behavioral theories, and machine learning.
Financial firms have a large stake in using social media networks. They want to deliver information to users quickly and efficiently. Firms want to be the trendsetters of their niche, and to be the center of the scattered information posted on the numerous social media platforms.
Although companies want to have an extensive social media presence, it does not always translate to more profits. Revenue from social media is not easily earned nor easily tracked. Increased traffic and interest on a company’s Facebook or X account, for instance, does not automatically translate to increased revenue, but mining social media data can provide valuable information. Companies mine social media tracking data to better understand users’ spending habits and product desires. Companies can then use what they learn to better target their online marketing to customers who are more likely prospects.
Social media mining can also show politicians the political leanings and attitudes of users, based on demographic information. It can give city officials a better insight into their citizens’ priorities and demands. By mining social networks, social scientists can examine either collective behavior (the behavior of groups of people) or individual behavior in order to better understand—and also predict—the behavioral patterns of people.
Social media mining is an emerging discipline, as is big data analysis. While analyzing and interpreting the data from social media mining has numerous implications, it can also be overwhelming because of the immensity of the data to be gathered. Nonetheless, these disciplines continue to grow and improve. By the 2020s, the accuracy and efficiency of social media mining had improved significantly, primarily though the use of deep learning and natural language processing—techniques that could provide a better understanding of large volumes of unstructured data from various social media platforms. Furthermore, a trend toward real-time analysis allowed for more immediate insights.
Bibliography
Amor Perez, Ester Maria. “How to Manage Complexity and Realize the Value of Big Data.” IBM, 28 May 2020, www.ibm.com/think/insights/how-to-manage-complexity-and-realize-the-value-of-big-data. Accessed 3 Dec. 2024.
Corley, Courtney D., et al. "Text and Structural Data Mining of Influenza Mentions in Web and Social Media." Intl. Jour. of Environmental Research and Public Health, vol. 7, no. 2, 2010, pp. 596–615.
Ganis, Matt, and Avinash Kohirkar. Social Media Analytics: Techniques and Insights for Extracting Business Value out of Social Media. IBM, 2016.
Gottfried, Jeffrey. “Americans’ Social Media Use.” Pew Research Center, 31 Jan. 2024, www.pewresearch.org/internet/2024/01/31/americans-social-media-use/. Accessed 3 Dec. 2024.
Gundecha, Pritam, and Huan Liu. "Mining Social Media: A Brief Introduction." INFORMS Tutorials in Operations Research. INFORMS, 14 Oct. 2014. doi.org/10.1287/educ.1120.0105. Accessed 3 Dec. 2024.
Kennedy, Helen. Post, Mine, Repeat: Social Media Data Mining Becomes Ordinary. Palgrave Macmillan, 2016.
Naone, Erica. "When Social Media Mining Gets It Wrong." MIT Technology Review. MIT Technology Rev., 9 Aug. 2011, www.technologyreview.com/2011/08/09/87637/when-social-media-mining-gets-it-wrong/. Accessed 3 Dec. 2024.
Rogow, Geoffrey. "In Social Media Mining, Some Go Slow." Wall Street Journal: Money Beat. Dow Jones, 1 Dec. 2014. www.wsj.com/articles/BL-MBB-30315. Accessed 3 Dec. 2024.
Tang, Jiliang, Yi Chang, and Huan Liu. "Mining Social Media with Social Theories: A Survey." ACM SIGKDD Explorations Newsletter, vol. 15, no. 2, 2013, pp. 20–29. doi.org/10.1145/2641190.2641195. Accessed 3 Dec. 2024.
Thakur, Nirmalya. "Social Media Mining and Analysis: A Brief Review of Recent Challenges." Information, vol. 14, no. 9, 2023, p. 484. MDPI, www.mdpi.com/2078-2489/14/9/484. Accessed 13 Dec. 2024.
"Top 15 Most Popular Social Networking Sites." EBizMBA Guide. EBizMBA, May 2016. www.ebizmba.com/directory/. Accessed 3 Dec. 2024.
Zafarani, Reza, and Mohammad Ali Abbasi. Social Media Mining: An Introduction. Cambridge UP, 2014.