This article provides an introduction to the world of big data to sociologists. It talks about the possibilities of using big data to identify new phenomena in human behavior. It also goes over the potential pitfalls of relying on big data and makes the case that big data works best when used in combination with other methods rather than in isolation. It concludes with what the future could hold for using big data for Sociology.
This article is an appropriate bookend for this semester. It goes over the major themes that we covered in detail in our classes and provides a good summary of the things we’ve learned. On noticing that this article was published in a Sociology journal, I was reminded that despite all the computing related things we’re doing, the ultimate goal is to further the study of humans in this connected world. All the machine learning and data analysis was a means to an end, which is obvious in retrospect but is not really in the forefront of your mind when you are deep in the throes of writing code.
The authors made an interesting point about scientists fixating on single platforms like Twitter, making the comparison to ‘model organisms’ in Biology. Behavior on Twitter is exclusive to Twitter and doesn’t necessarily reflect the wide range of human behavior. But even within so-called model organisms, there is such a rich research potential for studying human behavior. I don’t believe that findings need to be generalizable to be useful. Interesting observations can be made on single platforms and there is no need to constantly consider how generalizable the information is. Consider Finstagrams [1], a phenomenon where users are creating secondary accounts to be viewable by only select people. Regular Instagram accounts are often curated to be perfect and public facing. Finstagrams provide an outlet where users can just ‘be themselves’. I believe this could be a fascinating study, looking at what causes a user to make a Finstagram account, when did they first start appearing, what are the real world analogues, and so on. Looking at single platforms exclusively should not be dismissed for lack of generalizability for sometimes it is that lack of generalizability that makes findings interesting.
I think there is an interesting future ahead for this combination of Sociology and Computer Science. With everything that is happening with Facebook and the problems that arise from social media in general, I think this field holds the future in figuring out how to help solve the problem of humans in the online world, just like Sociology is trying to solve the problem of humans in the real world. It warrants keeping an eye out and seeing where things go from here.
[1] https://medium.com/bits-pixels/finstagram-the-instagram-revolution-737999d40014
