Phillip Brooker is an interdisciplinary researcher in the field of social media analytics, with a background in sociology and sociological research methods. Phillip co-convenes the Programming-as-Social-Science (PaSS) network (www.jiscmail.ac.uk/PaSS) which explores computer programming as a subject and methodological tool for social research and teaching. Phillip is also our social science expert and course instructor on Introduction to Python for Social Scientists. We spoke to him about his background in computational social science and what he’s been working on recently.
All of my work has been methodological driven really; the kind of overarching interest I have across all my research has been to develop innovative methods to support research involving digital (and sometimes 'Big') data of various kinds, primarily social media.
One of the first things I worked on in this area was the development of a social media (Twitter) analytics tool called Chorus, which offers an interface for collecting and visualising Twitter data in various different ways - see a recent paper of mine, Doing social media analytics for a more detailed discussion of the types of data collection that Chorus works with, and what you can do with them.
Working on this paper, we developed this idea of transposing a particular computer/information science methodology, 'visual analytics' (Thomas and Cook, 2005), as a framework for doing social science. At its core, a visual analytic approach to social science means not seeing visualisations as the end-result of a piece of research, but as analytic tools in and of themselves. This means that the visualisation needs to be seen as a tool not just for explaining data, but for deriving research questions, for pointing out methodological issues, for understanding the ethical implications of your research, for suggesting improvements in how you sort and classify data, and so on.
Since developing that kind of methodological thinking about digital/social media data, I've had the chance to apply it in various studies which use those ideas, but which keep methods at the forefront of my mind. For example, I have a forthcoming paper which analyses text comments left on a Guardian article discussing obesity as a public health concern, where we use visualisations of the data as a jumping off point to undertake a more qualitative analysis.
Recently, I've become more heavily involved with the idea of computer programming as a way of doing and thinking about social science. The same sorts of concerns apply here as with Chorus - the idea that we can use tools to not only do research but also to think about the research process, especially that which involves digital data. So currently I'm co-convening a mailing list on Programming-as-Social-Science, I've helped develop material for Introduction to Python for Social Scientists, and I'm currently writing a textbook on "Programming with Python for Social Scientists". The textbook is designed not only to teach social scientists how to code, but how to think about the impact and roles of things like software, technology, algorithms, code and data in our research.
I find this kind of thing really interesting - getting under the hood of the research process to examine the bits of it we don't often get chance to see, and also thinking more to the future about what social science could look like and do if programming skills were to feature in our core methods training.