Upon stepping off the plane from Canada, I was fortunate enough to attend the Oxford Internet Institute’s latest conference: Internet, Politics, Policy 2012: Big Data, Big Challenges.
It was a fascinating gathering of the leading experts in this small but substantially growing field. Since my background is in qualitative methods and I have yet to jump into web-specific methods, I felt it was a very good crash course on the collection and analysis of large-scale online data in application to many diverse areas. I’ve posted my conference notes for anyone who would like a gander. Also, we were promised that the presentation slides and webcasts would follow on the conference site. The papers for all the research have already been posted if you would like to access a goldmine of big data research and also fill in the gaps in my notes.
Points that stood out most to me were:
- The difficulty of defining ‘big data’. There does not seem to be a universal definition and it is also a term that depends on the technology you are using along with the subject matter. Some studies gathered millions of tweets while others collected a couple 100 threads from discussion boards. Some drew data over months or years while others siphoned comments from intense online dialogue lasting only a few hours. So while some researchers shrugged and said, “Well, I don’t know if this is big data, but…” it seems that the definition really is context-specific and that these studies were all ‘big’ in terms of their particular areas of research.
- The importance of forming interdisciplinary research teams to examine online phenomena. The unique nature of online data is such that the perspectives, methods and tools of computer science, the social sciences and even other ‘hard’ sciences (e.g. neural network modelling experiment-based psychology) all come into play. If a subject is only being assessed from one discipline, it is likely the researchers will lack either the tools to carry out the study or the framework from which to examine it.
- The utility and potential of big data. Given the nature of big data, it was often difficult for researchers to conclude any sort of causality between variables but they were still able to identify correlations and interesting patterns. This leads me to believe that big data can be a useful starting point to identify where researchers can drill down with other methods (e.g. surveys, interviews) to understand various layers of social phenomena. In this way, big data is the sort of exploratory research that helps sift through the noise of the web to signal to researchers where there may be something socially interesting for further investigation.
*The title of the post is a spin on the old Mr. Big chocolate bar slogan, “When you’re this big, they call you Mr.” However, I could not find the commercial on YouTube so I leave you with this gem:
Stefanie Duguay 