Data Usage in Social Media

Big Data at Berkeley

Published in

Big Data at Berkeley

10 min readDec 24, 2020

By: Jasmine Wang and Eddie Liu

Social Media in the 21st Century

As one of the world’s fastest-growing phenomena, social media’s impact ranges from simple communication between users to driving modern marketing campaigns to even influencing the ways in which humanity behaves. A significant reason contributing to the rise of social media is the rapid proliferation of internet access globally. Within the global population of 7.7 billion, approximately 3.5 billion people have internet access.

Taking a look at the history of social media’s growth, MySpace was the first platform to reach a million monthly users, doing so in 2004. Since then, Facebook has become the global leader in social media usage, followed by YouTube, Instagram, and WeChat respectively. The social media landscape is ever-changing, with some platforms rising to the top and others losing relevance altogether. For example, TikTok has seen unprecedented growth as of late whereas giants such as Hi5, MySpace, and Friendster have seen little to no share of the social media market since 2012.

As competition in the social media environment continues to intensify, functionalities such as video and image support have become mainstream and almost necessary components for a platform’s survival. This fluid development has distinct implications on different demographics of social media users. For example, 25–29 year olds prefer traditional platforms such as YouTube and Facebook while 18–24 year olds have proven to adapt more quickly to new platforms such as Instagram and Snapchat.

Another factor contributing to the growing popularity of social media is the advancement of technology in the form of personal electronic devices. In 2018, the average American user engaged with digital media for 3.6 hours on mobile, 2 hours on desktop/laptop, and 0.7 hours on other devices per day. This accumulates to an astonishing 6.3 hours of digital engagement per user every day! In fact, the percentage of U.S. adults who consistently use social media has increased from 5 percent in 2005 to a remarkable 79 percent in 2019!

This accelerated expansion of social media throughout the world rivals trends associated with revolutionary technologies such as computers, smartphones, and even the internet itself.

As such, many questions associated with the contemporary surge of social media will arise. How will the data collected from social media platforms change human perspectives? How will organizations use said data to draw insights and further their goals? Is social media data collection ethical in terms of privacy and user rights?

Social Data: What, Why and How

If I were to ask you, “What word pops into your mind when you think of ‘social media’?”, how would you respond? Perhaps, you’ll say “entertainment”, “fun”, “addicting”, and rightfully so. There is, however, another side that most tend to forget, and that is social media is also “informative”, “insightful” and most importantly “data”. Everytime you like a post, share a link or tweet a message, you are actually leaving traces of valuable data for public use. This is what we call “social data”. Social data is, at its core, raw information and insights formulated from individuals’ social media activities. Metrics include, and are not limited to, likes, comments, conversions, time spent, mentions, and even clicks.

Now that you know what social data is, you may be asking why it is even important to begin with? Although to consumers, social data may seem menial, it is an absolutely essential tool for businesses to gauge the interests of their customers, especially in a technologically savvy era. Have you ever wondered why the shoes you looked at when online shopping all of a sudden pop up on your instagram feed? Or why ads change depending on what region you travel to? This is the work of companies engaging with social data and using learned insights to determine their next step. In the first example, the shoe company sees that you placed their product in your wishlist and are reacting by feeding you ads of the item to increase the chances of a purchase. Similarly, businesses see that purchase trends differ on the east coast compared to the west coast, and thus adapt their advertising and marketing techniques accordingly. Both of these instances exemplify how companies react to consumers directly interacting with their platform.

Using a technique called social listening, businesses can also gain insight on current trends using other social media platforms other than their own. How? Let’s talk about ice cream. Most people attribute ice cream to a hot sunny day. As a result, it is only natural for ice cream businesses to funnel marketing and advertising resources towards hotter seasons. Utilizing a social listening tool, Ben and Jerry’s discovered that there were unusual spikes in ice cream sales on rainy days as well. Delving through mentions on social media platforms such as Twitter and Instagram, Ben and Jerry’s found that people enjoy eating ice cream on rainy days while watching Netflix. The popular ice cream company responded immediately and introduced “Netflix and Chill’d”. By “listening” to people’s comments and conversations on social media, companies like Ben and Jerry’s, are able to take full advantage of data to pioneer their businesses in new directions. That, in my opinion, is incredible.

So far we’ve only explored how social data benefits business-to-consumer companies (B2C) by examining questions like, “How should I adjust my marketing strategy to appeal to a specific age group?” or “What factors contribute to customers’ buying trends and how can I take advantage of that to maximize revenue?”. Besides B2C companies, social data is also incredibly important for business-to-business companies (B2B) as well. Social data answers questions like “What do consumers’ enjoy in a competitor’s brand that I do not have?” or when selling to a business “What does this CEO or CTO value and how can I market myself in a way that is appealing to his or her taste?”

Notably, there are a plethora of applications that come with social data, and hopefully, I convinced you of the importance of social media in current day commerce.

Case Studies

In this section, we will delve into specific examples of how large versus small companies utilize social data to improve their businesses.

Mastercard

“Mastercard is a multinational financial services company that operates payment processing networks to connect consumers, merchants, banks and more businesses in more than 210 countries and territories.” (Commpro) You most likely have seen, at one point, the Mastercard logo at the corner of a debit or credit card.

With the advancement of technology, payment methods have evolved from cash to card to now mobile payments, with the latter streamlined by Mastercard. The company created the Conversation Suite — a “dynamic, global insights and engagement engine” — engineered and managed by social experts to monitor and analyze conversations around the world in real-time. Listening to conversations on social media platforms (such as Twitter, Instagram, Facebook, etc …), they became the first financial company to successfully use social data as the primary means to refine mobile payment.

Through extensive social listening and data cleaning/analysis using multi-language search-string algorithms and human-validated analytics, Mastercard determined the two main consumer concerns to be lack of acceptance and data security. They continue to utilize the Conversation Suite to refine said concerns becoming the leading company in the mobile payment industry by 2015. In 2015, Mastercard boasted a 95% customer satisfaction rate, an increase of 18% percent since 2013.

Spotify

If you are a Spotify user, you are probably familiar with the “Discover Weekly” feature. This feature provides listeners a list of custom songs every week according to that individual’s music taste. Using methods like collaborative filtering, natural language processing and audio models, Spotify is able to cluster songs according to how listeners interact with the platform. At its core, Spotify uses social data to predict the listening trends of their consumers in order to provide better service to their listeners!

Small business

In his TedTalk, data scientist at Irish Water, Barry James, talks about a meaningful experience he had when staying at a small hotel with his family during a vacation. His kids were throwing an uncontrollable tantrum in the lobby and a worker, knowing the family’s room number, 220, offered to deliver breakfast cereal and toast directly to their room. This he argues, is where small businesses have a data edge over larger businesses. They are able to provide personalized care in the right way, at the right time. While unpersonalized care is not necessarily a bad thing, James argues that personalization is not feasible for a larger company and thus smaller companies should take full advantage of it. Because of the accessibility of social data, any company, despite its size can adapt it in a way that is beneficial and impactful to their targeted consumer.

The Ethics of Social Data Consumption

As companies accumulate increasing amounts of social data, the usage of said data begins to draw public scrutiny regarding the ethicality of such collection and manipulation of data. In 2014, contractors and employees of Cambridge Analytica, a British political consulting firm, stepped into the limelight for selling psychological profiles of American voters to political campaigns. As tens of millions of Facebook users had their data manipulated to influence hundreds of elections globally, backlash struck and news about the Facebook — Cambridge Analytica data scandal spread worldwide.

As knowledge about social data usage multiplies, many moral and ethical questions come into view. One such question asks: is the social data being collected considered private or public information? Social media users nearly always agree to a set of terms and conditions for each platform that they use, but oftentimes users pay little to no attention to formalities such as these. The online setting makes the line between private and public extremely ambiguous. For example, how should researchers approach a public post in a private Facebook group? What about trending vs. non-trending hashtags on Twitter?

A common argument for the proliferation of social data usage is the concept of informed consent. In traditional research, participants must sign forms within surveys and questionnaires acknowledging the use of their responses. However, in social media research, participants’ data is accessed without explicit informed consent: researchers can retrieve their data almost instantaneously. Since most social media users do not properly read the terms and conditions of each and every platform they have an account on, most are unaware of their participation in studies. On top of this, the scaling of large data set populations makes it nearly impossible to get informed consent from every participant.

Another facet of consent is its revocability. One’s consent to use their personal data should be inherently revocable, but ambiguity arises when delineating whether or not social data has been truly dismantled. If a user deletes their post or account, are they renouncing all access to their social data? If companies continue to store their data, are users aware of this? Consent should innately be revocable, but the act of withdrawing on social media remains unclear.

Anonymity within big data is yet another critical issue when it comes to the ethicality of social data usage. When publishing or even analyzing individual data extracts, such as a Tweet, it becomes increasingly difficult to protect the identity of unknowing participants. The original data extract can be easily associated with the original poster, which clearly pinpoints a single user. In addition to this, image-, audio-, and video-based data poses severe challenges to anonymization as vocal and facial features are handily attributed to individuals. Lastly, as people and organizations identify users to some degree for marketing purposes, metadata must be searchable in order to separate and filter. All of these pressing concerns pose the question: how can companies adequately anonymize data for secondary usage?

Conclusion

Ultimately, the swift growth in popularity of social media usage brings about unparalleled potential for individuals and groups to capitalize on social data for more informed decision-making. However, along with this opportunity comes many moral and ethical question marks as well. As social media only continues to weave itself into the fabric of our society, we must ask ourselves, at what point does the usage of social data to draw insights become too invasive and what can we as data researchers do to protect the privacy of social media users?

Sources

Feel free to reach out to us if you have any feedback, or if you want to know more about the major. Also, follow us on Instagram @bigdata.berkeley and visit our website at bd.berkeley.edu if you want to learn more about Big Data at Berkeley!