As some of you know, last week I went to Amsterdam with a group of fellow students (and some academics of course) to meet with other students from the University there and discuss BIG THINGS.
We flew from Southampton airport, and I still can’t quite believe that it takes less time to get there than it does to get to London. The hotel room was amazing – the picture doesn’t really do it justice (or reveal the full horror, depending on your point of view). The toilet and shower were effectively glass pods within the room itself, complete with coloured mood lighting, with the bed taking the whole of the far end of the room, under the window. We could probably all have slept in it, it was so huge.
This was Wednesday evening, work started on Thursday morning. Various topics for the two-day workshop were put forward a couple of weeks beforehand. Fortunately, we weren’t expected to produce anything specific, which would have been very difficult because the BIG THINGS were very big indeed. Before this, we had plenty of time to introduce ourselves to our colleagues who, like us, came from all different parts of the world, and talk about our research. Honestly, the breadth of topics being studied under the banner ‘Web Science’ is astonishing, and this is what makes interdisciplinarity so important and interesting. We had linguists, computer scientists, sociologists, psychologists…. all brought together because their research centres around some aspect or affordance of the web. After lunch, we separated into groups to discuss the BIG THINGS.
This was the topic for our group: What characterises news stories that first emerge through social media (rather than through the official news media). I was working with a colleague from Southampton, and two colleagues from Amsterdam, and we spend all of our allocated time on the first day just unpicking this. Questions arose such as ‘what is ‘the news’?’ What makes a story worth repeating i.e. what are ‘news values’? Are the values, first suggested in 1965 by Galtung and Ruge, still relevant in the digital age we live in now? What do we mean by ‘social media’? Twitter? Facebook? Comments sections of news stories posted online by the traditional media? (after all, you often see whole conversations happening beneath particularly contentious stories). Then there’s the BIGGER THING around how you even find news stories in social media in the first place. Ok, all the traditional media outlets have their own accounts, as do the non-traditionals like Buzzfeed. And we know that some stories naturally emerge in social media first because they involve people using their smartphones to record events as they unfold, like the recent bombings. But what else is there? The real problem is that you won’t know if a story has originated in social media until it breaks in the traditional media, and then you’re working backwards through time. How do you find it first?
I found this whole discussion, and the strategy that emerged from it, immensely satisfying. First of all, I was able to draw from my own background in media studies. It’s always heartening when you have some experience and knowledge you can bring to a group to help it on its way. Secondly, the topic carried on from something I became interested in when I was working with colleagues in China. There, we were using a data set of tweets from Twitter that focused on the general election in India. The research students I was working with – all from China, Singapore or Korea – were entirely focused on writing code to extract information from the data (directed by the group leader, which probably shouldn’t have happened, but hey-ho). This was fine, but we ended up duplicating work that had already been done, and learned nothing new. What would have been more interesting would have been to see how the various stories regarding each of the candidates/parties bounced between the traditional news media and was retweeted and/or added to by the ordinary people using Twitter. It would have been much more interesting to find out how this happens in a network that has formed around an event like a general election, and to find a way of modelling it, preferably interactively.
Finally, it was thrilling on a personal level to know that I understood the theory behind the tools we were suggesting might be used to both identify news stories in social media, and pick up new ones as they emerged. With only 140 characters to deploy to create words (at least, in English), identifying a cluster of words that suggest a news story centering around topic ‘x ‘ is appearing is challenging to say the least – and its perfectly possible that there is no way of identifying emerging news stories successfully. You just won’t know until you try. The two colleagues I was working with from the University of Amsterdam had plenty of experience with the kind of analysis that could be applied to our problem, but as good interdisciplinarians were more than able to discuss the wider aspects, such as the limitations of cluster and sentiment analysis when sites like Twitter shorten words and use slang.
In the end, we were able to do no more than resume our discussions on Friday, and formalise them into a short presentation and an illustrative poster. My wish now is that we can meet again (or meet with other interested parties) and turn our discussion into something more formal, perhaps something fit to submit to a conference. Certainly the topic as it stands is enough to fill an entire PhD.
Anyway, back to Thursday evening when this happpened…. yes, a boat trip around the canals, combined with dinner. Lovely!
I’ll post more photos in a separate gallery, as we took advantage of the weekend and instead of flying back on Saturday, flew back on Sunday evening instead. Much fun was had by all…. some more than others. *cough* Coffee shops *cough*. Not me, obviously. Others.
And finally, let me thank the Web Science Institute at the University of Southampton and the equivalent at the University of Amsterdam for an absolutely brilliant, thoroughly positive and very useful exchange. Thank you!