Big Data and Cloud Culture - To the Cloud: Big Data in a Turbulent World - page 197

and that, when necessary, qualitative states can be rendered qualitatively.
For example, a quantitative content analysis of search terms relating to
lu provided Google with what it believed was a string of terms that cor-
related with lu outbreaks, thereby enabling researchers to predict, earlier
than ever, the spread of lu. If, on the other hand, one chose to carry out
a big-data analysis of a subjective state, say by associating positive Twitter
posts about the Toyota Prius with sales of the car, then one might assign
numerical values to capture the strength of responder posts. Or big data
might run an analysis that combines the results of numerous customer-
satisfaction surveys that assign a number to each possible response, such
as a 5 for strong dislike or a 3 for simply disagreeing with a statement.
After all, strongly like or dislike represents a more powerful attraction than
just like or dislike . The measurement of quantity is not only central; it is
absolutely essential to the transformative capacity of big data. As two of
its proponents attest, “Just as the Internet radically changed the world
by adding communications to computers, so too will big data change
fundamental aspects of life by giving it a quantitative dimension it never
had before” (Mayer-Schönberger and Cukier 2013, 12). There is much
to be said for quantitative analysis. It renders complex behavior, as well as
mental states, easy to process and analyze. It is no wonder that big-data
specialists believe that “the more quantitative it is, the better” (Morozov
2013b, 232). The ease of analysis, the opportunity to draw broad gener-
alizations and then to make predictions, provides a strong temptation to
reduce all methodological approaches to quantitative ones. Indeed, the
hot new profession of data scientist knows only quantitative approaches.
Moreover, big data makes it possible to avoid the need to sample a popu-
lation, and all of the risks associated with accurately representing a larger
group, by examining results for an entire population.
The problems with relying solely or primarily on quantitative analysis
are today more often than not ignored, but that is a mistake. Quantita-
tive research provides a scientiic gloss on behavioral or attitudinal data
that is often far messier than the numbers make it appear. Social scientists
are well aware of the limitations of working with data on reports of law-
breaking behavior that are often massively skewed by the human limita-
tions of witnesses, police, and the vagaries of plea-bargaining and trials.
Nevertheless, big-data supporters and their corporate sponsors continue
to press for what is euphemistically called “predictive policing” (Bachner