Everywhere I look, digital humanities and social sciences are increasingly in vogue and utilizing big data in that research is apparently the next big thing. Big data has been recognized as great resource for academics also, and behavioral sciences are increasingly required in policy making *again/finally*.
I have a bit of a past studying fairly large and multiple online datasets from social science angle. This is probably why I was offered a PhD student position at University of Helsinki Consumer Research Centre – an offer that I happily embraced! These last few weeks, I’ve been busy filling my knowledge gaps on theory and fashion of “what is hot in academia”. After my stint in the industry and SMEs I’m more pragmatic than academic, and that needs balancing. Luckily my experience gives me an edge; it “only” needs to be translated into academia concepts and vocabulary. To that end, I am currently going through Lauri Eloranta’s neat course introducing me and my fellow Helsinki students to computational social sciences.
In addition to studying these interesting university courses and reading outstanding books like John Galbraith Affluent Society, I am working for the Citizen Mindscapes project led by Docent Mika Pantzar (PI) and senior researcher Krista Lagus. The goal of this project is to bring large datasets and the opportunities in analysing them to the reach of University researchers all over. To advance the project, I did a quick test using Finnish academic data analysis platform Korp, interface to Finnish data and representing our Finnish effort towards Open Science. Korp is developed originally for linguists, however I understand that its usage is not always as straightforward for people with social sciences background.
I jumped at the chance to put on my pilot suit, run a test and wrote a test report on potential social sciences big data -newbie user stories and cases, because here at Consumer Research Centre one of the best things is that we are expected to develop new tools and understanding to aid the whole social sciences community at University of Helsinki! So, I wrote down what I did and aimed for in my mock-up start-of-research “semi-quanti” study, and quickly analysed my pilot run. I was delighted to notice I already used several items in Lauri Eloranta’s list of data mining methods – namely, summarization, classification, clustering and anomaly detection, in addition to the general “exploration of data and tool”.
I’m more than convinced that analysing big data indeed is the way to go for social sciences. For any scientist out there with interest yet feeling of helplessness, the argument is simply: using big data and analytical platforms to analyse discussion data is not a totally different thing from analysing texts like a pile of interviews or open-ended survey results. The tools have changed and introduced a few more layers to consider as confounders but also as sources of extra information. But the phases in the process should be recognizable!