“Data Sciences and Society” Reading Group
DISCUSSION NOTES BY SAM GOMES:
We began our fall 2019 programming on September 11 with a Data Sciences and Society reading group meeting led by Brice Ménard from the Department of Physics and Astronomy at Johns Hopkins University. During his presentation, “Invisible Worlds Revealed by Data,” Ménard demonstrated the increasing prevalence and utility of the long tail distribution in economic analysis.
Ménard eased us into his talk about the long tail by asking us how we understand numbers and quantities. He walked us through the different ways that we visually process numbers to reveal to us the connection between our vision and cognition. As a group, we discussed the limits to our accuracy when estimating quantities. Comparing the examples of one versus ten dollars and ten million versus one hundred million dollars, we observed that amounts we can no longer grasp seem closer together than amounts we can understand. Professor of Anthropology Veena Das, a PI of the Seminar, related these occurrences to accidents, earthquakes, and lottery winnings: all of these things seem far away until they happen to someone close to us. Together, the reading group agreed that scale, time, and chance affect our cognitive ability to understand range and closeness. Ménard used this observation to conclude that visualizing the long tail, when it is used to model the infinite options in a variety of markets, is beyond humans’ unassisted cognitive capacity. Some participants dwelled on what exactly qualified as cognitive capacity (and failure thereof), seeing as we can comprehend the long tail when it is explained to us. This led to Ménard specifying his understanding of the term as one of visual thinking or “grasping”—as opposed to “mathematical thinking” which kicks in when the more immediate forms of cognition fail.
In Ménard’s words, the world became saturated with consumption options in entertainment about twenty years ago, when more data, faster technology, and more storage allowed for the exposure of more content to consumers. The long tail distribution, which represents all of the choices consumers make, is an increasingly occurring phenomenon as a result. Ménard also observed that studies of the long tail have massively proliferated across sciences in the recent years as well.
To prepare for the gathering, reading group attendees read the first and eighth chapters of The Long Tail: Why the Future of Business Is Selling Less of More by Wired magazine’s previous editor-in-chief Chris Anderson. These chapters discuss the changing nature and geography of these markets brought by fringe sales. In Chapter 1, Anderson writes that digital services with infinite shelf space can offer niche products and profit from them with “the same (or better) margins as a hit” (24). With popularity no longer having a monopoly on profitability, the “new shape of culture and commerce” is a long tail (Online music popularity, 25). Chapter 8 introduces readers to linguist George Zipf, who found that industrial processes, population statistics, and word use trends can be observed with “power law” distributions and see a long tail when graphed. Anderson then discusses the model’s applicability to sales when carrying capacity is no longer a problem and shelf space is unlimited. The long tail captures the abundance of distribution and choice, challenging principles of scarcity in traditional economics.
Ménard told us that this ungraspable variety of selection that Anderson considers “infinite” is a source of the increased worry we feel about the control we have over our choices. Das pointed out that surveillance mechanisms and other apparatuses reinforce our distrust of technology and contribute to this unease. Some participants pushed back that grasping was a matter of habit, but Ménard held that habit was unnecessary for a certain window of cognition. There was also some contention around whether the long tail is a new phenomenon, or whether we are noticing it more now that we are more aware of it.
To end his presentation, Ménard compared the availability of data to wetness: the way a tree extracts water in a desert is the way a website extracts data from users. It is also the way a user extracts music, entertainment, or information from a streaming source or search engine. “Tree cords” extract these small bits of information, and just as the roots of a tree work to accumulate five liters of water per day to feed the tree, tree cords come to gather large amounts of information.
During the question-and-answer portion of the talk, we unpacked the language of and concepts related to the long tail. Canay Özden-Schilling pointed out that Anderson seems to shroud the ideas of data and shelf space in wonder, forgetting that data still has a material component to it and shelf space is not necessarily “infinite” as Anderson suggests. Some of us found that there was a lack of discussion in Anderson’s book on labor. It was interesting to learn about how vendors like Amazon and eBay are relating producers to consumers in a new way, but little was said about how this process affects producers and the production of content. We also noted that Anderson’s account did not make room for monopolistic platforms’ and algorithmic recommendation systems’ role in constricting (and not elevating) consumer activity.
The long tail has altered approaches to connecting with consumers and optimizing consumers’ exposure to data. While there still needs to be more discussion concerning content creation and the division of labor under this new era of long tail economics, there was great value in coming to understand how practitioners view this new method and its increasing prevalence. To reduce some of the stress that comes with being online, Ménard proposed that if consumers consider themselves the tree gathering data and reclaim this idea of abundance for themselves, they can optimize the time and energy they spend looking for entertainment, media, resources, and information. Although much about technology is left out of our control, he suggested, we can take personal advantage of what it allows us to enhance our experiences online.