I am a first year PhD student and my interest lies in the use of Artificial Intelligence techniques in classroom education. The primary basis of AI techniques such as machine learning is data. Some of the things I am hoping to understand in my research are who chooses to collect a certain type of data and why, and what are the factors influencing the data collected by educational software present in classrooms. This course seems to be a great way to build a solid understanding of what we mean when we speak of data, and how the nature and purpose of education is influenced by the large-scale, automated data collection and analysis enabled by various educational software.
What I am also hoping to get through this course is a deeper understanding of why this matters. Historically, data has always been collected on students — whether that is exam scores, attendances, absences — and is used in making decisions; it is the foundation of how education works. Analyzing trends and progress, and making distinctions between learners (for better or worse) is not new. So what is different with the advent of “Big Data”? What changes when data is collected on the scale that it currently is through computational techniques? While on a surface level, I feel like I know the answer to this question, I am hoping that by engaging with the readings and activities in this course I will be able to establish a clear theoretical basis for what changes with the digitization of student data, and its consumption by AI technologies.
Just thinking about the visualization exercise (which I have not gotten to yet), I’ve been quite overwhelmed by the enormity of what it means to have to collect the kind of data collected in Dear Data. That is to say, if I were to have to create a visualization of how many times I checked the time during the day, it would mean I actually have to be present enough to catch myself when I look at the clock (instead of it being a subconscious activity). Personally, over the past several years, I have found being truly present quite the struggle. I am bombarded with so much data on an everyday basis, that data almost has no meaning, or rather, I am forced to remove the data from my life so it does not overwhelm me. An example would be the smart watch I owned a few years ago. The smart watch’s reading of my heart rate and constantly telling me to “calm down” made me so anxious, I am pretty certain my heart rate went up every time I wore it in anticipation of the fact that it was going to tell me by heart was beating too fast. What was meant to help me take better care of my health, had a very real adverse impact on it.
There is something about digitized data that is both useful and useless all at once, and something about how data is captured and used that transforms the experience of the data. I remember using a wellness app that would tell me how much time I had spent on my phone that particular day, broken down by apps. An example looked something like:
Gmail : 18 times, 40 minutes
Whatsapp: 48 times, 120 minutes
Facebook: 30 times, 6 minutes
Aside from various inferences of me as a user from these statistics, I found myself increasingly overwhelmed by the end of the day statistics. Sure, it meant I used my phone a little too much and suffered a fair bit from open-scroll-close syndrome with social media. But what did it really mean?? I eventually deleted the wellness app and consciously decided to tackle facebook usage alone. I tried to become aware of every time I opened the app, and in the early days of this attempt, I would only notice after I had been scrolling for ten seconds that I had even opened the app. The more I noticed, the faster I would catch myself in the process of using the app or wanting to click on the app icon. It was exactly the same data (technically) that the wellness app provided, but the experience of the data, physically, mentally was completely different.
The Dear Data visualizations have really got me thinking about how data in itself is nothing until it is captured and experienced in some way. The visuals in the book make the data almost a source of art, where I’m first drawn to the illustration first and what the data represents second. At this current point of time, this morphing nature of data depending on how its captured and presented is what I’m chewing on and finding intriguing.
One response to “Introductory Post”
Excellent reflection Meenakshi. I especially like your comments on feeling bombarded by data, and the difficulties of both noticing and recording ‘events’ for the data visualization task. One of the reasons we ask for this to be done manually–both in the recording of data and its translation into a visualization–is to highlight how data collection and its graphical display are normally enacted in ‘black boxed’ systems that we can rarely get ‘inside’. A data processing machine has no problems about being bombarded by data. Manually collecting data and hand drawing visualizations, we believe, helps foreground the complex processes and practices that are increasingly allocated to computers, code, software and algorithms (which are themselves located in organizations, whether commercial, nonprofit or governmental, with their own aims and objectives around data). Gillian Rose and colleagues have a nice study on data visualizations as ‘interfacial sites’, by which they mean that a visualization is a complex coming together of data collection methods, aggregated datasets, analytics software and algorithms for sorting the data, graphical software for its representation, as well as data analysts, designers, human-computer interaction professionals and so on. In our case, we’re performing all those roles, and surfacing the ways that data and visualizations have to be made.