In my eyes, data analysis is about numbers and graphs that tell you a little about a lot of people. Without interacting with a single person you can sweep the internet using a variety of programs (including TAGS, which I talked about in a previous post) and categorise anyone by interests, friends, education, etc.
“Twitter’s success has made it a rich research site for scholars interested in online interaction, information dissemination, activism, and a plethora of other subjects. The sheer volume of users, tweets, and hashtags has made the site a favourite for quantitative data analysis and “big data” number-crunching.” (Marwick, 2014, p. 109)
The fact that you can get so much data so easily is amazing (and a bit scary) but it’s only really scraping the surface. In-depth research is needed to understand the people behind the statistics.
In ‘Twitter and society: an introduction’ (chapter 9) Alice E. Marwick, an academic writer, looks at how other ethnographic research methods were also important in her research of Twitter. She looks at ‘in-person’ field work, such as interviewing, as well as observing Twitter users individually online.
I’m not surprised by her findings at all, which is why I can understand why you have to use different research methods. Contradictions and inconsistencies in data mean you need to take a closer look to find out the bigger picture.
“Comparisons of a person’s discussions of Twitter with their Twitter stream can reveal an added layer of useful information. For example, the information gathered by researchers in face-to-face settings may be consistent, or divergent, from the uses demonstrated by collected tweets or the type of information considered proper to share. […] Reviewing tweets about an event where ethnographic data was gathered can help flesh out participants’ meaning-making practices about their activities.”(Marwick, 2014, p. 115)
Marwick’s experience using Twitter is also quite similar to my own. Sometimes you can use the search bar and find nothing. You type in something very specific to find things unrelated to what you actually want.
Also, when contacting potential interviewees directly on Twitter, the less followers they have, the more likely they are to reply to you. I have found with customer-facing companies that if you include them in one of your tweets first they are more likely to reply when you contact them directly. But then they prefer the actual interview to be done over email, rather than on Twitter itself.
I’ve always preferred interviewing to data analysis. But now I’m looking into the big world of social media more closely I’m realising that for each research project I am going to have to find the balance between quantitative research and qualitative research.
Marwick, A. E. 2014.Ethnographic and Qualitative Research on Twitter. In: Bruns, A. et al. eds. Twitter and Society. New York: Peter Lang Publishing. pp. 109-121.