Over the last few years, the idea has taken hold that “big data” is driving far-reaching, and typically positive, change. “How Big Data Changes the Banking Industry,” “Big Data Is Transforming Medicine,” and “How Big Data Can Improve Manufacturing,” are characteristic headlines. “Big data” has become ubiquitous, powering everything from models of climate change to the advertisements sent to Web searchers.

Even in a society in which acronyms and sound-bites pass for knowledge, this familiar formulation stands out as vacuous. It offers us a reified name rather than an explanation of what the name means. What is the phenomenon denoted by “big data”? Why and when did it emerge? How is “it” changing things? Which things, in particular, are being changed – as opposed to merely being hyped? And last, but hardly least, are these changes desirable and, if so, for whom?

Big data is usually defined as data sets that are so large and complex – both structured and unstructured – that they challenge existing forms of statistical analysis. For instance, Google alone processes more than 40 thousand search queries every second, which equates to 3.5 billion in a day and 1.2 trillion searches per year;[1] every minute, Facebook users post 31.25 million message and views 2.77 milion video, 347,222 tweets are generated; by the year 2020, 1.8 megabytes of new information is expected to be created every second for every person on the planet.[2]

The compounding production of data – “datafication,” in one account [3] – is tied to proliferating arrays of digital sensors and probes, embedded in diverse arcs of practice. New means of storing, processing, and analyzing these data are the needed complement.

A quick etymological search finds that the term “big data” began to circulate during the years just before and after 2000.[4] Its deployments than quickened; but this seemingly sharp-edged transition into what Andrejevic and Burdon call a “sensor society”[5] actually possesses a deeper-rooted history.

The uses of statistics in prediction and control have long been entrenched, and have increased rapidly throughout the last century[6] – as is pointed out by a working group on “Historicizing Big Data” established at the Max Planck Institute for the History of Science. The group emphasizes that big data must not be stripped out of “a Cold War political economy,” in that “many of the precursors to 21st century data sciences began as national security or military projects in the Big Science era of the 1950s and 1960s.”[7]