Yesterday, CLT ran for the second time a workshop exploring the use of social media AS research data (as opposed to using it as tools TO DO research). We first ran this in June 2013 as an experimental, exploratory workshop, which was a great success, but this time we wanted to shorten and improve it based on the feedback we received. By all accounts, it was a great success. I say “by all accounts”, because as misfortune would have it, I was struck with illness and could not make an appearance, after weeks of organising and preparing for it. Unfair beyond belief – especially since my colleagues raved about to me how good everyone’s presentations were.
The programme mixed group discussions with short presentations. Unfortunately, I missed the two new case studies by Phil Brooker (Brunel) and Nick Anstead (notes on last year’s studies here). Phil’s study outlined his analysis of the #notracist hashtag, using Chorus data collection and analytics software. Phil says that the #notracist hashtag operates quite differently from event concerned hashtags, whose spikes reflect times, trends and contexts. The #notracist hashtag is used steadily, by largely unconnected set of tweeters – yet after deeper analysis it appears that routine patterns do emerge. Phil is a member of the Chorus team and blogs about the software on the Chorus tumblr page. This presentation was the one I had really looked forward to, and so I continue to curse the gods for their untimely striking me down with malady. Phil recommends you take a look at the introductory video on the site first if you are interested in using Chorus and his work with Dr Sanjay Sharma is written up on the Dark Matter Wiki.
Ella McPherson then introduced the room to tricky ethical aspects of social media research, anchoring her views around a core statement by danah boyd, that “just because it is accessible, doesn’t mean using it is ethical”. In other words, don’t fall into forgetting that social media data is connected to actual human beings. In the case of politically sensitive situations, re-publishing and using that data may for example serve to identify the original posters, and actual harm may be caused to them. This raises security and anonymity issues, questions about consent and the spectre of power relations. I remember Ella’s presentation from last year and can honestly say it is excellent. Her slides give some insight into that fact.
Finally, Nick Anstead from the Department of Media and Communications expounded on his own work using Twitter, analysing the #bbcqt hashtag, and then explored the opportunities (large data sets e.g.!) and challenges (large data sets, e.g.!) of using Twitter. According to my colleagues, this was really a great introduction to the field of doing research with or on social media data, explaining what can and cannot be done – for the participants in the room who weren’t sure about the topic, though clearly interested, Nick gave them a good sense of how to start their own research, what to expect and to look out for if you want to see his slides.
Last year we started a google doc collecting some good online tools and links for participants and those who missed it. We have also included the links to the four cases studies we didn’t have time to read in this document. The Google doc is open to anyone to edit, so please feel free to add any new tools you want to share with us.