It was a little difficult trying to understand the exact message of this paper without knowing the context. I assume it was written as a counterpoint to the idea that Big Data is a wonderful and irreplaceable new tool. Maybe that idea was particularly strong and well argued, and there was a need for its critique but for me most of the points made in the paper looked obvious. Yes, when working with data, one needs to carefully consider its origin and what exactly it does and does not represent. Yes, most data needs pre-processing and that pre-processing might inject subjectivity into one’s analysis. Yes, using public or semi-public data for research without its authors’ explicit consent is sometimes ethically questionable. But the way I think about Big Data already takes these things into account. To me it’s just a single tool that can be used in social research along with interviews, polls and other more standard data gathering methods. When writing a paper that uses survey data as evidence, the authors are expected to comment on why they believe the data is accurate and representative, both qualitatively and using statistical arguments. The same should be true for Big Data.

The thing I found interesting in this paper was the idea that, since using Big Data usually requires technical expertise, there is now a shift in what training researchers need to have and how this affects their perspective. While I can’t comment on the academia side of things, I have definitely noticed a growing culture among programmers/developers/hackers of working on small side-projects that use modern technologies to reflect on social sciences/art/humanities - ‘the real world’. An example of this is a project that uses machine learning and natural language processing to auto-generate a politician’s speech based on the transcripts of all their previous speeches. Projects exploring political speeches would traditional have been done by the academia. Instead, this was done by an independend developer in their spare time and published in Medium, a blogging platform. As someone who’s a part of the hacker culture, I see this as the democratization of humanities. However, I can definitely understand why people traditionally trained in humanities would find this suspicious - most programmers’ knowledge of social theory and analysis is as amateurish as most academians’ knowledge of deep learning techniques.