Issue link: https://cp.revolio.com/i/457819
22 | THE DATA CENTER JOURNAL www.datacenterjournal.com privacy is DeaD a ccording to a Pew Research Center poll from late 2014, the vast majority of users are aware of and concerned about a lack of privacy as well as the use (or misuse) of their personal information by both the private and public sectors. e results indicate that 80% of social-networking users "are concerned about third parties like adver- tisers or businesses accessing the data they share on these sites," and 70% "are at least somewhat concerned about the govern- ment accessing some of the information... without their knowledge." Yet even though 91% (effectively, everyone) believe that consumers no longer have a say in how their information is collected and used by companies, use of social media is still growing. e only explanation is that despite complaints about privacy, most users see more value in the services that social- media providers offer. In a sense, that fact is not surprising: although some individu- als may use Facebook, Twitter and/or other social-networking sites just as a means of communicating with possibly distant friends and family, many treat them as a platform to proclaim their accomplish- ments, views and all-around praiseworthi- ness. Even scrupulous users can fall prey to this attitude. Psychology professor Jean Twenge of San Diego State University said in a New York Times piece, "Many studies find that narcissists have more friends on Facebook and post more, especially provocative material. ey also post more oen on Twitter. Apparently, narcissists thrive on social media...Narcissism clearly leads to more social media use, social media use leads to positive self-views, and people who need a self-esteem boost turn to social media. It is less clear whether social media directly causes narcissism, at least in the short term." Regardless of the precise relationship, social networks at least enable narcissism—and perhaps encourage it. And since narcissism and privacy don't mix, it's unsurprising that despite sensing the danger, users find it difficult to leave a medium that feeds their self-image. is platform for self-expression has led to a treasure trove of information. Its almost drug-like effects—causing users to ignore the warning signs just to post one more selfie—have bypassed the inhibi- tions that have prevented governments and companies from collecting such personal and detailed information by other means. Regardless of the effects on users, however, what's to be done with all that data? mining DiamonDs or Lumps of coaL? On the surface, the biggest challenge to gleaning useful information (how- ever that's defined) from social-media repositories is addressing all the so-called unstructured data: information that isn't easily segmented in a nice spreadsheet or database fashion. at includes pictures, videos, text (to a lesser extent) and so on. Breaking down all that data into classifiable (not necessarily even useful) information requires extensive computing power, but machines are less than adept at taking into account the vagaries of human behavior. Depending on the intent of the analyst, even "correct" conclusions may still be wrong in either the moral or business sense. A quintessential example of the potential danger of data mining is a well- known case in which department store Target analyzed customer purchases to attempt to identify prime candidates for marketing maternity-related products. Unfortunately, the company's analysis was a little too good: it sent advertisements to a teen whose parents were unaware she was pregnant. is sort of situation, although slightly different from the context of social media, shows the kind of faux pas that can result from attempting to categorize people by their data without taking the human factor into account. But such situations revolve more around what's considered socially accept- able and not so much what is actually true—Target wasn't incorrect; it was just inconsiderate. Even ostensibly anonymous data can create disturbingly detailed pic- tures of a person's private life, as a study by Stanford University researchers discovered. Looking just at so-called metadata from mobile-phone usage of several volun- teers, the researchers found that "phone metadata is unambiguously sensitive, even in a small population and over a short time window. We were able to infer medical conditions, firearm ownership, and more, using solely phone metadata." Social- network data, being much more detailed, could likewise be highly informative even when stripped of names, IP addresses and similar identifying information. Nevertheless, data mining can lead to erroneous results. Even a relatively high success factor for a given metric can lead to increasing uncertainty as a larger picture is built. For instance, say an analysis is Social networks have done what the Stasi could only dream of, barely falling short of even the omnipotent state in George Orwell's 1984. In collecting untold masses of information about billions of people (who have supplied it willingly), social-media providers have at their virtual fingertips potentially powerful databases. Efforts are underway across a number of industries to glean valuable conclusions from all this information, but the question is whether the rewards are worth the costs.