The Subjectivity of Data

One of the best things about doing a PhD is being able to ‘audit’ classes. This means I can participate in just about any subject I wish, but I’m not bound to complete assessments and the like. Talk about bi-winning.

One of the classes I’m auditing this semester is The Art of Scientific Visualisation. The first ‘lecture’ (it’s an online subject, so more of a reading than a true lecture) was a bit of a history lesson looking at the evolution of scientific illustration. Its transformation from an artistic pursuit around the 16th century into an attempt for objectivism by the 19th century is an interesting story. What struck me most was the consistency of data being presented in the style of the day.

I want to suggest that the images have an agency of their own and that even the most intricate renderings are an artistic construction literally embodied with interpretation. In short, they are not a faithful reproduction but a snapshot of the period in which they were made. This “snapshot” is not a linear one but rather a thematically based approach to ideas and discoveries.

The above snippet from the lecture sums this idea up well. Some of the original anatomical illustrations were created by combining artificial poses (popular of the day) with observational studies from cadavers. Why this was the case could be debated, but it seems consistant – even today – that scientific representation employs a degree of artistic license, even when trying to avoid it.

When working with others to create visual expressions of data, layers of subjectivity and interpretation are compounded:

Unable to draw, Leeuwenhoek employed the services of skilled draughtsmen – in a sense it was a process of collaboration and hence again interpretation of what was seen through his microscope. Apparently the draughtsmen he employed would often have debates with him about how to ‘accurately’ render his findings, peering over his shoulder as he observed the invisible.

My interest in this subject is born from the heavily graphic world of data visualisation. A passing look at information aesthetics will tell you that not even brutally quantifiable statistics will result in objectivity. Presentation requires personal interpretation and subjectivity that goes with choosing to focus on one aspect, whilst omitting another. It’s a discussion that will need to be explored in my own research, and an interesting one at that.