Beyond Techniques, Ethics: The Lessons We Need Today

I started interpreting before I had a degree for it. I worked in retail and some friends needed to know what was happening for meetings. And so, like any good friend, I raised up my hands and tried to let them know what was going on in the meeting. They occasionally asked me to go other places too and, like a good friend, I went.

Except, in the long run, I may not have been such a good friend. In the United States, we have the Americans with Disabilities Act, as well as the Rehabilitation Act of 1973. These laws are designed to set in place ideas like “reasonable accommodations” and fair access. Bringing a friend, while comfortable at times, doesn’t guarantee fair access and can get very uncomfortable in places like the doctor’s office. Add in my own lack of knowledge and training on navigating certain cultural differences, and you have yourself a very unqualified interpreter.

Yet, my story is nearly universal among those who came into the community the way I did. We often get tapped to provide access to spaces that are inaccessible or hostile in their worst iterations. Long term, we fight for access. This story, too, shares a familiar pattern.

Technicians in a Trade

Many of us start visualizing data in much the same way that I started interpreting. We learn to make charts, to follow best practices, and to spend a ton of time on accuracy. We work on color, formatting, and on refining storytelling techniques. We become downright magical in our ability to surface patterns or find a predictable behavior. We evolve to be rock solid in our techniques at connecting the dots in so many ways but one: the effects downstream.

In 2014, I applied to be a consultant to a firm based out of Cincinnati. I was provided the Ohio Schools data set to do a test visualization and made an exploration that looked at demographics and then flowed into a workflow for finding the best school for your child. It looked like this:

I spent a ton of time on visualizing, while I struggled with shaping the data. That affected the analysis. From a technical side, I worked hard on the branding. The technician model would appreciate this.

I very much wanted to make a tool where a parent could find a good fit for their child (in my example, gifted female). The technician model recognizes one space where I’ve gone awry: averaging the averages. Yet, it misses something bigger. Have you spotted it yet?

Perhaps, a different visualization, one by Candra McRae, will help call out the larger problem.

As highlighted by Candra, school segregation is a continuing problem. Hover over Ohio and you see we have 17 districts with open desegregation orders. Additionally, we see 57% of US adults still believe segregation remains a moderate to serious problem, and the cluster of OCR complaints helps further showcase the depth of this issue. As Candra notes in her viz, these complaints are voluntary fix it orders and lack real teeth to create change.

Revisit the snapshots above of my dashboard and you can see how my work might serve as a tool for perpetuating segregation, rather than reducing it. Harm is often served on a platter of good intentions and my viz is no exception. This can also happen when we visualize sensitive topics, such as death and suicide (content warning).

Shifting Paradigms

When I started working in data back in the late 2000’s, the data world was undergoing a radical transformation. We were amassing more and more data and slowly learning to integrate our various sources. Fast forward to today, and it’s nearly impossible to truly de-identify data at scale.

We went from guessing in the past to gaming, our swaths of data quickly becoming part and parcel with nearly all of our experiences. Your data from Facebook, for example, can flow throughout your online experiences, being used to guide you towards very specific purchases and towards various beliefs.

Psychologically, we like to believe our personalities and preferences are fixed. Rather, they’re extremely malleable to settings, place, and culture that surround us. Robert Sapolsky has done a deep dive on how we define us and them, and how easy it is to manipulate these group identities. We see this when we go to college (GO BUCKS), or on vacation (yes, I really will wear this outfit when I get home…hey, why is this buried this deep in my closet?) and when we move to very different areas or even jobs (I have always been the person that would dye my hair green).

Data further complicates these biases. Seemingly invisibly, we generate numbers without often considering how the number came to be in the first place. Some things may be obvious or even laughable after the fact, such as the artist who pulled a wagon with 99 phones through the streets of Berlin, causing the streets to clear due to “traffic” while others – such a pothole reporting – are less so. Further, technology is often projected as unbiased, when again and again, we see it’s dangerously the opposite.

It is no longer safe to focus solely on techniques and on becoming better technicians. As creators, we will always see the good we can create and fuzzy over the harm without proper training. And, if anything, you’ve seen even with proper training, we can still get it wrong. Yes, even me as shown above. Ethics work best in groups.

Professionals, Not Technicians

While the technician focuses on skills and techniques, the professional model centers on using skills and techniques in an ethical fashion. It takes those same focuses, but channels them towards value sets, which often include doing no harm as a first tenet.

We saw this discussion occur in the medical industry in the early 1900’s when medical devices were unchecked and ungoverned. It happened in chemistry with the development on the H-bomb. And, it’s happening now with data ethics.

These changes are hard and fraught. They force us as practitioners to slow down and ask if we should create something. One of the best resources I’ve seen thus far comes from EthicalOS.org. You can download their checklist and dive deeper with their toolkit.

We are past the point where technique is enough. It’s time we acknowledge our roles as professionals and practice accordingly. Just like with refining our technique, learning ethics takes continual time and effort. It’s worth it – not just for better skills – but a better world.