Following last week’s overview this post provides a bit more detail on the doing of social media analysis. This course ran for three weeks, each week focusing on a different aspect and using a different tool: understanding & gathering (TAGS), analysing (Tableau), and visualising with social network analysis(Gephi). As stated, the course does work for a total beginner, I didn’t need any understanding of coding, although working knowledge of Twitter and excel was helpful.
It is possible to engage with the material at one of three levels, each with an increasing time commitment. For 2 hours a week, where all you do is read the course content as presented, a total beginner could develop a very superficial understanding of the terms and concepts used in social media analytics. Four to six hours would be enough to attempt all the core practical exercises and to select one or two pieces of additional reading per week. This is what I did, and I came away with a basic understanding, I felt comfortable using the tools and, by referring back to the course instructions, would be able to do some very simple analysis on my own (identify who is talking about a topic on Twitter, how they are connected to each other etc.). To really integrate social media analytics into a research project or evaluation would involve deep reading of all the additional material, completing all the follow on practical exercises and practicing with the tools using my own data (and without step-by-step instruction). This required more time than I had, and is something I would only commit to if I could work with an expert as part of a live project.
In week 1 we used TAGS, a free Google Sheet template which lets you setup and run automated collection of search results from Twitter, developed as a hobby project by Martin Hawksey.
TAGS is easy to use, you type in your search terms, the time period you want to search for and you can set it to update automatically, to capture new tweets as they are published. If you were working on a subject likely to be covered on Twitter, it might be useful to run this sort of data collection at the start of the project, concurrently with an initial literature or Google search.
The skills for using TAGS will be familiar if you have been involved in identifying search terms for literature reviews – understanding context is important, trial and error helps fine tune and so on (as I was typing this I had a sudden flashback to a research project from a few years ago. The initial literature review included academic papers on newts – we were searching in relation to young people not in employment or education!).
The course suggested some questions to help with focus and understanding context:
- What is the topic, issue or event you’re researching?
- What groups are involved in this topic, issue or event?
- What keywords and/or hashtags are associated with it?
- What things do you hope to or predict you’ll find?
- Create a question to sum up your ideas.
My first attempt looked for tweets on the role of social media in social research. Even after tweaking the way I searched, I only came up with 30 tweets or so. While this told me something – the topic isn’t much talked about on Twitter– I decided to try something more mainstream. So I searched for tweets to address the question ‘What is More United UK and who is interested in it?’
I think this generated about 700 tweets, lots of which were re-tweets. It was still quite a new topic and there was nothing to see when looking for trends in the tweets. To get a meaningful answer to ‘who is interested in More United UK’ I would need more information than I could access through Twitter alone, for example demographics. Even then the data would be extremely biased, limited only to those who were discussing the new movement on Twitter, rather than on Facebook, in the press, and between themselves in day to day life.