The current swirl of debate around big data and challenges of mining, capturing, curating, storing, and analyzing large data sets shows little sign of abating. In any given day, most workers will generate, use, provide, interact with, and interpret some kind of digitally-rendered data. Ways of working—and everyday living and learning—are increasingly infused with information and decisions made elsewhere and informed by big data analytics.
Some of this digital data is knowable and accessible by workers and citizens but much is not. More extensive data flows are now occurring between machines, with humans only one of many actors in these processes (Hayles, 2006). The growing datafication of professional work is evident in how it is distributed across crowdsourced data and predictive analytics; bots that automate online tasks; and new regimes of accountability and surveillance implicit in many digital interactions.
Gray (2016) describes datafication as “ways of seeing and engaging with the world by means of digital data” (para 3). Digital data refers to the things of professional practice that have been digitized: converted to binary numbers.
To examine this phenomenon, ProPEL Matters is hosting a series of three blog posts.
Big data is sometimes pitted against small data: constructing a new binary of value and credibility. However, Kitchin (2014) observes that “big data should complement small data, not replace them” (para 9), suggesting that there is a need for sophisticated qualitative work in big data research.
But there is more here for academic researchers (especially qualitative researchers) to do beyond merely supplementing big data work. These are the issues we will explore in our blog series that will unfold throughout June and July 2017:
- Anna Wilson will suggest a way out of the big-small data binary: a third way in which big and small data are recognized as connected and co-constituting. In this third way social researchers can explore for themselves new possibilities for presenting and analysing educational data.
- Cate Watson will push back on the preoccupation with size (aka “big” data) suggesting that this diminishes the value of the insights to be gained from in-depth examination of a single case. Small data can be really useful and she will explain why.
The data we draw on emanates from the oPEN (Online Professional Education Network) project focused on the development of a 16-week online masters course. The interactions between people and digital things that unfold in online learning environments, such as BlackBoard, generate digital traces that are easily recorded and stored. But in this course, the digital data was deliberately subjected to more assertive learning analytics. Such analytics, enacted with social network analysis and visualization software, represent a form of datafication: a purposeful datafication of the work of both teaching and research.
The instructors in this course could have relied on the simplistic and clunky analytics in the LMS. Or, like many online instructors, selectively ignore them, instead forming their own judgments based on their interactions with students. If the instructors in this study had turned to the “Retention Centre” in their LMS (which generates analytics of a sort) they would have encountered an obstreperous digital “thing”. There is no easy way for the average digitally literate instructor to see, question, or change the algorithms that generate the data in the LMS and its assemblages of charts, tables, and warnings.
And so, in this project, various data traces and patterns in the online space were made visible through the use of Social Network Analysis (SNA) software: NodeXL. Although we enrolled this more advanced software, and a dedicated analyst, this study highlights a range of analytic options that are doable in everyday teaching—and research—practices. See Adams & Thompson (2016) for an account of how this analytic work unfolded.
Illustrated is how the growing swaths of trace and archival data online open up new ways to analyze and visualize data that often then informs professional actions, decisions, and responsibilities. What meaningful insights can such analysis generate? In this project, the teaching-research group was uncertain about what was knowable and how we (or more accurately we-what) knew it: uncertain about how such analytic work can constructively inform professional work.
Looking more closely at how such work is performed helps to question who-what is doing this work. In the swirl of datafication practices, who-what is being datafied? How are responsibilities and decisions being reconfigured as work is increasingly outsourced and delegated to digital devices?
Datafication is a messy process despite the work that goes into trying to create order. At the very least those researching and teaching in digital spaces are looking at a re-distribution of labour between human actors and their digital counterparts. These are questions I will be exploring in my paper at the upcoming ProPEL conference (Thompson, 2017).
References
Adams, C., & Thompson, T. L. (2016). Researching a posthuman world: Interviews with digital objects. London: Palgrave Macmillan. doi: 10.1057/978-1-137-57162-5.
Gray, J. (2016). Datafication and democracy: Recalibrating digital information systems to address societal interests. Juncture, 23(3). Retrieved from: http://www.ippr.org/juncture/datafication-and-democracy
Hayles, N. K. (2006). Unfinished work: From cyborg to cognisphere. Theory, Culture & Society, 23(7-8), 159-166. doi: 10.1177/0263276406069229
Kitchin, R. (2014, June 6). Rob Kitchin: “Big data should complement small data, not replace them.”
Thompson, T. L. (June, 2017). Digital data and professional practices: A posthuman exploration of new responsibilities and tensions. Paper to be presented at 3rd ProPEL International Conference (Linköping, Sweden).