This article, surprisingly originating from a sociology journal, seems to succinctly talk about everything we spoke about this semester in our class. They speak about big data’s enormous use in understanding human behaviour. Especially behaviours/behaviour patterns that are difficult to record or are often misreported in self-reports.
As is fit by a sociology journal, they cast the world of big data into various manifestations – digital life, digital traces, digitalized life and instrumentation of human behaviour. When the authors say that digital life can be viewed as generalizable microcosms of society, I wonder if this is appropriate. Given the fluid and incremental way in which platforms grow, does it seem appropriate to give and keep a label?
As we’ve seen several times before in this class, one worries about the ethical conundrum in mining big data to gauge insights in sociology. While the tracking of phones of several students to understand the ties between friends (from Facebook), the students were aware of this. We previously read a paper about Facebook (without the knowledge of its users) tweaking its feed to gain knowledge about emotional reactions to the posts on their feed.
We are also surrounding ourselves with gadgets and objects that are constantly providing information that can be exploited possibly for things we would not consent to. In a sense, we are creating our own surveillance state. Dr. Michael Nelson, during a recent seminar here, spoke about this very problem. By using fitness tracking devices, virtual assistants, we are unknowingly and unwittingly providing valuable data for analysis. The fitness app Strava faced some flak for this when US soldiers in Afghanistan used the app for tracking while running around army bases there. The soldiers unbeknownst to themselves were clearly giving away locations of secret army bases. Obviously, we are still having some trouble keeping all the data being misused.
Another issue with big data that was mentioned by the authors that seems prevalent is the issue of inclusivity. During the talk given by Dr. Rajan Vaish, I remember his mentioning that even during their study (Whodunit?), they found it difficult to involve the rural population in India. Ofcourse this is in part due to the expensive, metered internet connections but this is also a question of interest (in case of other major online platforms) for the community.
The authors themselves have outlined several other issues that come with big data analytics in sociology. One of the more familiar ones is the issue of bots, puppets and manipulation. We have, over the course of this course, seen several ways we can curb this behaviour making the data that is available to us more meaningful. But this problem is present for now skewing a lot of the ongoing analyses.
Finally the authors talk about qualitative approaches to big data and I almost laughed out loud! We were battling this issue with a much much much smaller data set until even a few days ago. The authors promise computationally enabled qualitative analyses that will help us analyse data that is beyond the capacity of armies of grad students to read. To this we say thank you!
This article was a fitting way to end the semester and the course. It was in itself a summary of sorts making it slightly difficult to summarise and reflect on.
