Reflection on “Big Data, New Epistemologies and Paradigm Shifts” by Rob Kitchen

With the rise of big data, the world is entering a new petabyte age which will rapidly transform the lens from which we view science and study human behavior. Rob Kitchen in his article, “Big Data, New Epistemologies and Paradigm Shifts”, discusses how this transformation will challenge already established epistemologies in the humanities, social sciences, and science in general. Chris Anderson addresses this topic as well in an article, “The End of Theory: The Data Deluge Makes the Scientific Method Obsolete.” Anderson expands off of a quote by George Box who says “All models are wrong, but some are useful.” With a massive amount of data just at our fingertips, it will naturally become a valuable solution when faced with questions or problems. Simply run the algorithms, look at the patterns, and make observations based on any correlations that are seen. J Craig Venter used this approach and went from observing singular organisms to entire ecosystems.

This method is taking over the way in which we conduct research/science and advancing the measurements we use to collect information. Some biological models are now seen as flawed, the scientific method could potentially be ended, and models such as those in quantum mechanics will be seen as obsolete. As we collect more data and information, a working model to explain all these things becomes less plausible, unless we turn to data. The problem here lies in the fact that data will only infer a correlation, but not explain causality. Hence, the way we make measurements will need to change for the adaption of both data and models working together. While the advent of big data is advancing the revolution of measurement, the revolution in measurement is also helping to elevate data science as an emerging field of study. A new system of measurement is forcing the field of data analytics to be reframed to fit this mold. Thus, we are entering a fourth paradigm focusing on data centered exploration through approaches in empiricism or data driven science (with inductive, deductive, and abductive methods).

Because the humanities and social sciences are less tangible, Kitchen is uncertain how data will impact these fields. A lot of people argue that the arts and literature can not be treated as data because it’s not measurable the same way physical sciences are. Contenders also argue that humans are too complex for their nature to be deduced by measured laws citing examples such as wars, violence, racism…etc. While this is true, data deluge is advancing a better understanding of humans and these behaviors. Kitchen discusses how data used in digital sciences is descriptive (looks for patterns), whereas data in social sciences has to be used for causality. Thus, he concludes that big data analytics must be employed under a different epistemological framework that makes reasonings/infers causality when observing social sciences. Part of the favor of big data over strictly models is that data can extract additional insights. Knowing this, data can offer a more holistic view when studying complex systems such as the environmental system discussed in Rob Kitchens paper. Regarding behavior, big data offers proliferation of the digitization of unstructured data. It also offers numerous opportunities about historical, political, economic, social, and cultural data. All of which will be valuable insights towards improved understanding of human development as a complex and adapting social/economic system.