April 2021 – Data Visualisation

Final Visualisation Blog Summary

Considering data from the perspective of learning, teaching, and governance has been a useful exercise, in which I have reflected on the collection, analysis, and presentation of data in education. In the first block, I was reminded that data in isolation may not demonstrate learning, even though sometimes some data is better than no data (Brown, 2020). In the second block, it became clear that the datafication and commodification of education is changing the role of the teacher (Williamson et al, 2020). Lastly, in the third block, the power of data was apparent as the governance perspective shed light on the “questions of power” in relation to the type of data collected, how it is understood and communicated, and for what purposes (Anagnostopolous et. al., 2013: 7).

In the first block, I tracked music habits, Twitter notifications, phone usage, emails and questions. Each visualisation was bound by time, highlighting the passing of time through data in a way that became unexpectedly personal. When comparing the visualisations, I was reminded that “what counts as education when it comes to digital data is what can be counted” (Williamson, 2017: 46). For example, counting emails is simple; however, tracking emotional engagement with Twitter is complex because it can be fluid, and is not easily bound by time. The question is if either demonstrates learning, and if so, if one is preferred over the other. Working in technology, it is a good reminder that data is personal, needs context, and that not everything can be counted.

Furthermore, there should be careful consideration of the type of data collected and the purpose as certain data impacts data privacy (Bulger, 2016). As technology usage increases, data privacy concerns will become increasingly complex, because artificial intelligence and other technologies can be used to track a student’s every move. For example, facial recognition can be used to read and understand student emotions (Chan, 2021), and wearables augment the type of data points available about the human body while learning (Knox et al, 2019). In my personal data reflection in Week 4, I analysed the data collected by my iPhone. While I was pleased to see that I spent 2 hours exercising in the dashboard for my “quantified self” (Eynon, 2015), the phone only tracks the app usage time, not the fitness value without the integration of a wearable technology. From a learning perspective, this highlights that further reflection is needed to assess if the “simple act of using numbers” does indeed demonstrate learning, or simply highlights that something happened as the 2 hour block shows in my iPhone dashboard (Eynon, 2015).

In the second block, I tracked sleep, emotions, and distractions under the assumption that these impact student engagement. Behavioral data could be used for gamification and personalised, or adaptive, learning if artificial intelligence or wearable technology was integrated into learning platforms (van Dijck et al, 2018); however, these are not data tracked by learning platforms today and raise data privacy concerns. In my technology experience, engagement is tracked by mouse clicks, time, comments, etc. From a teaching perspective, this highlights the importance of selecting valuable data points because a dashboard can limit the view that a teacher has, and in turn, impact their perspective of students (Williamson et al, 2020). For example, my visualisations provide a limited view of what impacts my ability to engage. Additionally, the dashboard could unknowingly limit teaching methods rather than positively impact them (Brown, 2020). As highlighted by Bulger (2016: 4), in classrooms, teachers leverage learner-centered instruction and personlise teaching based on “interpersonal cues…. subject matter expertise… knowledge of how people learn, and knowledge of each student, to determine individual needs, adjusting their lessons in response to questions and behaviors.” In the remote classroom, this is not as easily accomplished. The teacher needs dashboards to bridge the gap both from a learning and teaching perspective.

Similarly, this limited perspective may be transferred to teachers, if using the data for performance purposes because the data can become “proxy measures of the performance of staff, courses, schools, and institutions as a whole” (Williamson et al, 2020: 354). A distinction is needed between the data collected to demonstrate learning and the data collected to demonstrate teaching methods. This is an interesting consideration when remembering that the data actors are not always educators, but technology companies and other non-governmental organisations (Williamson, 2017). From a technology perspective, more data is an easy upsell, which translates to additional revenue and happy shareholders. From a teaching perspective, more data is not always beneficial when teachers may lack necessary skills to analyse dashboards and recognise bias in an algorithm that produced the dashboard (Brown, 2020).

In the third block, I tracked technology-enabled interactions, getting help, and anxiety with a focus on the purpose, value, and power of data through the governance perspective. The physical act of collecting data and creating the visualisation demonstrated all three as it became the beginning of understanding the complex layer of abstraction that influences governance, which in turn is pushed back down to teaching and learning. As such, learning, teaching, and governance become a cycle of, and for, data. How data is collected is invariably influenced by collectors (Ozga, 2015), but even more important, is acknowledging the number of actors (human and non-human) that interact with the data before it becomes a performance metric, or a report (Anagnostopolous et al., 2013). Technology improvements and actors like the ‘Big Five tech companies’ (van Dijck et al, 2018) have enabled the datafication and commodification of education, giving rise to ‘fast policy’ and influence over the education system (Williamson, 2017). This results in “questions of power” from the initial collection through dissemination of the data (Anagnostopolous et. al., 2013: 7).

In summary, a simplistic view of data in education is that it provides an opportunity to demonstrate learning, assign a value to teaching, and serve as insight or transparency for governance (Ozga, 2015). The visualisation task enabled a view into data for the purpose of learning, teaching, and governance, highlighting that this simplistic view is far from the truth. The data process – from collection to dissemination – most importantly, highlighted the separation of the student from the data and the risk of generalisation and unintended perspectives (Anagnostopolous et. al., 2013). Lastly, it continuously reinforced that what matters is what can be counted (Williamson, 2017), and ultimately, that data impacts “how we practice, value, and think about education” because it allows for the categorization of the good and the bad (Anagnostopolous et. al, 2013: 11).

Word Count: 1008 without citations, 1087 with citations

———————————————————————–

Sources

Anagnostopoulos, D., Rutledge, S.A. & Jacobsen, R. 2013. Introduction: Mapping the Information Infrastructure of Accountability. In, Anagnostopoulos, D., Rutledge, S.A. & Jacobsen, R. (Eds.) The Infrastructure of Accountability: Data use and the transformation of American education.

Brown, M. 2020. Seeing students at scale: how faculty in large lecture courses act upon learning analytics dashboard data. Teaching in Higher Education. 25(4), pp. 384-400

Bulger, M. 2016. Personalized Learning: The Conversations We’re Not Having. Data & Society working paper. Available: https://datasociety.net/pubs/ecl/PersonalizedLearning_primer_2016.pdf

Chan, Milly. (2021, February 21). This AI reads children’s emotions as they learn. CNN Business. Retrieved from https://edition.cnn.com

Eynon, R. 2015. The quantified self for learning: critical questions for education, Learning, Media and Technology, 40:4, pp.407-411, DOI: 10.1080/17439884.2015.1100797

Knox, J, Williamson, B & Bayne, S 2019, ‘Machine behaviourism: Future visions of “learnification” and “datafication” across humans and digital technologies‘, Learning, Media and Technology, 45(1), pp. 1-15.

Ozga, J. 2016. Trust in numbers? Digital Education Governance and the inspection process. European Educational Research Journal, 15(1) pp.69-81

van Dijck, J., Poell, T., & de Waal, M. 2018. Chapter 6: Education, In The Platform Society, Oxford University Press

Williamson, B. 2017. Digital Education Governance: political analytics, performativity and accountability. Chapter 4 in Big Data in Education: The digital future of learning, policy and practice. Sage.

Williamson, B. Bayne, S. Shay, S. 2020. The datafication of teaching in Higher Education: critical issues and perspectives. Teaching in Higher Education. 25(4), pp. 351-365.

During the ‘governing’ block, I created visualisations on technology-enabled interactions, getting help while injured, and anxiety sparked from three different categorisations of my life – personal, work, and school. The visualisations and readings focused my thoughts on three main themes with policymaking and governance in mind:

The purpose of data: What is purpose of the data and is the data ‘good’?
The context/value of data: How can the context or value of the data be better included in the outcome?
The power of data: Who holds the power of data?

What is purpose of the data and is the data ‘good’?

Policymaking and governance are reliant on data to provide insights, serve as evidence, and enhance transparency for the purpose of decision-making (Ozga, 2015). The ultimate goal is to know what is ‘good’ and what is ‘bad’, e.g. are the students learning, do they have the skills deemed necessary to advance, are the teachers effective, what schools are doing well, etc. Ironically, however, “Rather than empowering the people, the data may constrain what people know and how they think about their schools” (Anagnostopolous et. al., 2013). This is why getting to the answer of purpose and what is considered as ‘good’ data is important.

Note: the definition of ‘good’ for this discussion is what is useful and true. In an ideal scenario, this would also include data that does not infringe on someone’s privacy; however, certain private data points may be useful and true to the policymaking.

As a result of a ‘need to know’ culture and pressure to create policy and governance, the process appears to start with the end result rather than starting with the data and through analysis, finding an outcome regardless of ‘good’ or ‘bad’. While this is the scientific process, i.e. stating a hypothesis, making a prediction, and testing to determine the outcome, the readings give a sense that the process of iteration is limited when it comes to creating education policy. ‘Fast policy’ is the result of the increased number of actors (human and non-human) in education policymaking (Williamson, 2017). More data is being collected, enabled by the increased use of technology and improved infrastructure; however, the context seems to be forgotten as the game of telephone is played with the data after collection (Anagnostopolous et. al., 2013). The data collected travels through many actors and processes by the time it reaches those using it for policymaking and/or reporting. It is also invariably influenced by those doing the collecting (Ozga, 2015).

Williamson (2017) quotes Bowker (2008: 30), “A good citizen of the modern state is a citizen who can well be counted – along numerous dimensions, on demand.” This statement assumes all aspects of us as individuals can be quantified, yet this is not true. There are aspects of us as individuals that cannot be neatly quantified, defined, or categorised as evidenced by my own attempt to track anxiety. As a result, determining what ‘good’ data is, is complex and one that needs iteration and agility. ‘Fast policy’ and the use of technology may enable this iteration, if the assumption is that the policymakers are willing to be as agile and change existing policy as new information is available. The ideal for many would be that the data serves the education system (and it’s policymaking and governance) rather than a political or material purpose, which is often the case (Pinsloo 2020).

How can the context or value of the data be better included in the outcome?

Anagnostopolous et. al. (2013: 7) state, “Assigning students, teachers, and schools a numerical rating and classifying them as ‘failing’ or ‘effective’ says little about the causes of failure or what can or should be done about it and about school success and how to maintain it.” Context is important in understanding the data, but the context cannot always become a data point itself. For example, not all context is a quantifiable data point that can be added to, or understood by, a technology tool. Examples of this could include emotions and skills that are difficult to categorise neatly, like creativity and emotional intelligence.

In my own visulisations during this block, the context became key to understanding my own data as simply looking at the data points without knowing that I had, for example, been injured one week would dramatically change the interpretation and outcome. Imagine if the data was collected on a student, but the student was unable to provide that data point because it wasn’t possible in the system, or available as a question. The policy created from these data points, which become an indicator of performance, would likely not be ideal.

The statement made by Anagnostopolous et. al. aligns well to this: “As they define what kind of knowledge and ways of thinking matter and who counts as ‘good’ teachers, students, and schools, these performance metrics shape how we practice, value, and think about education” (2013: 11).

Who holds the power of Data?

The data that is now collected, is not only controlled only by government, but also non-governmental organisations like private sector companies (Williamson, 2017). These non-governmental organisations have increasing influence over education as they have a seat at the table to decide what can be inputted into the systems, the research that should be done, who (or what) completes the analysis of the data, and who will have access to the data.

Anagnostopolous et. al. (2013: 7) state, “Determining what kind of information about the nation’s students, teachers, and schools is collected and how it is processed, disseminated, and used, by whom, in what form, and for what purposes involve questions of power. They also reflect and privilege particular ways of thinking, particular values, and particular notions of worth.” What this highlights is that the student, the teacher, and the school that the data is collected on, no longer holds the power of their data. The power is held by the non-governmental organisations and governments who are analysing and reporting on the data. Similarly, this was a reason why I personally didn’t want to collect or highlight certain things in my own visualisations. As soon as the data has left my hands, the power to it has also left.

Taking the ‘infrastructural perspective’ approach (Anagnostopolous et. al., 2013), more time should be spent on identifying what data is collected for what purpose as well as how it is collected, and ultimately, pushed upstream to the end consumer. This large-scale datafication process involves countless actors (human and non-human), and the outcomes are now often readily available to those far beyond the school where it was collected (Williamson, 2017). Ultimately, there is a danger of a layer of abstraction as the data can become vague, or general, be interpreted from numerous perspectives, and end up being used in ways that were not originally intended (Anagnostopolous et. al., 2013). This is a key point when thinking about policymaking and governance in education. The hope, nonetheless, is that the policies and governance enacted benefit those in the education system, rather than limit or hinder them in any way.

————————————————-

Sources:

Ozga, J. 2016. Trust in numbers? Digital Education Governance and the inspection process. European Educational Research Journal, 15(1) pp.69-81

Prinsloo, P. 2020. Data frontiers and frontiers of power in (higher) education: a view of/from the Global South. Teaching in Higher Education, 25(4) pp.366-383

Williamson, B. Digital Education Governance: political analytics, performativity and accountability. Chapter 4 in Big Data in Education: The digital future of learning, policy and practice. Sage.