Big Data and the Little Big Bang: An Epistemological (R)evolution

Abstract

Starting from an analysis of frequently employed definitions of big data, it will be argued that, to overcome the intrinsic weaknesses of big data, it is more appropriate to define the object in relational terms. The excessive emphasis on volume and technological aspects of big data, derived from their current definitions, combined with neglected epistemological issues gave birth to an objectivistic rhetoric surrounding big data as implicitly neutral, omni-comprehensive, and theory-free. This rhetoric contradicts the empirical reality that embraces big data: (1) data collection is not neutral nor objective; (2) exhaustivity is a mathematical limit; and (3) interpretation and knowledge production remain both theoretically informed and subjective. Addressing these issues, big data will be interpreted as a methodological revolution carried over by evolutionary processes in technology and epistemology. By distinguishing between forms of nominal and actual access, we claim that big data promoted a new digital divide changing stakeholders, gatekeepers, and the basic rules of knowledge discovery by radically shaping the power dynamics involved in the processes of production and analysis of data.

Some Contemporary Aspects Related to the Big Data

“…[5]. While big data uses mathematical analysis, optimization, inductive statistics, and concepts from nonlinear system identification to infer laws (regressions, nonlinear relationships, and causal effects) from large sets of data with low information density to reveal relationships and dependencies, or to perform predictions of outcomes and behaviors [2]. In a broad sense, big data is referred to as a socio-economic phenomenon associated with the emergence of technological opportunities to analyze huge amounts of data [4].…”

Calculating the Souls of Black Folk

“…Indeed, the actuarial approach to human decision making 34 continuously reproduces a mythology 35 correlate to poor social outcomes as the cause of poor social outcomes. The objectivist rhetoric surrounding big data 36 directs us to adjudicate the downstream social impact of emerging technologies, asking questions like: Does the Allegheny Family Screening Tool (AFST) produce fair outcomes? Who do we hold accountable if it doesn't?…”

How can big data shape the field of non-religion studies? And why does it matter?

“…In recent years, the discussion about big data kept rapidly growing, not without a considerable level of hype among both scholars and business professionals. According to a popular-and strongly criticized [91][92][93] -narrative, big data came to revolutionize everything making previous science obsolete. 94,95 More recently, rather than suggesting replacing small data with big ones, researchers argued that integration between these two methodologies or ''cultures of modelling'' 96 is both possible and necessary.…”

“…An approach called ''co-labeling'' was developed to deal with ambiguous problems where labels of users in the training sample are uncertain. 106,107 Co-labeling is basically a multi-view learning method that combines classifiers trained on different views to improve the Privacy concerns 108 and the effective availability of usable data 91 to the researchers are potential obstacles to a research design relying on co-labeling to develop an efficient classification strategy of religious nones. Regarding this, further limitations, at least from a sociological standpoint, are certainly represented by potential sampling/selection bias due to the access to limited segments of the population (e.g., the ones present and active on social media platforms, the owners of smartphones) as well as the relatively limited time span accessible through this particular approach.…”