Why Data Science?

“Philosophy is an unusually stubborn attempt to think clearly” –William James

I’ve recently started a Data Science Fellowship at Flatiron School. If you’re not familiar with the term data science: it’s an emerging field that uses computer programming to extract insights from data sets, often very large ones.

Before this, I completed a Ph.D. in philosophy, specializing in emotions and rationality.

Emotions matter because, among other things, they help us to be attuned to things that are important for achieving our goals. When they’re inappropriate, though, they can lead us to make wildly, disasterously inaccurate conclusions. My research was on taking emotions into account in formulating norms of rationality. How do they impact our thinking? How should we manage them?

After graduate school, I did some teaching, published some papers, traveled to international research conferences, and (perhaps the thing I’m most proud of) helped launch a website that challenges philosophy faculty to teach nonstandard texts, topics, and methods, including philosophy from cultures
that are neglected by the Anglo-American canon that most of us are trained in.

So why data science? Why now? When I first took up Python programming, I was hooked. Lately, nothing makes me happier than working out solutions to problems with code.

This, of course, is essentially the same reason I love philosophy. A core part of the work of a philosopher is untangling some of the gnarliest tangles of logic there are.

Now I have an amazing opportunity to see what a philosopher can do with some Python packages and a bunch of data.

I’m still in the process of figuring out how to integrate my identity as a philosopher with this new endeavor, so I’m going to use this blog post to work through some questions I have about the connections between philosophy and science and the ways that I think my philosophical training prepared me to do data science well.

I’ll start with an influential (though not uncontroversial) way of thinking about the relationship. It originated in the Enlightenment period in a time when science was undergoing a major paradigm shift from scholasticism to Newtonian science. Before the shift, science was considered “natural philosophy” — a subset of philosophy. John Locke, an important philosopher from the Enlightenment period and a champion of the new science turned this on its head, describing philosophy as a “handmaiden” to science.

What this means, loosely speaking, is that philosophers work behind the scenes doing the dirty work that needs to be done in order for science to be at its best.

After all, when you’re formulating questions and hypotheses, interpreting data, developing theories and models, and applying evaluative standards you’re either doing philosophy or you’re helping yourself to theoretical assumptions that are philosophical in nature.

As “handmaiden” to the sciences, the philosopher’s task is, first and foremost, to clarify concepts and evaluative frameworks needed for scientific research.

Of course, as philosophers, we’re also interested in evaluating the institution of science itself. Does it improve our lives? Does it bring harm? (And does that matter?)

This is one of many good reasons not to get too carried away with the “handmaiden” metaphor.

For most of history, philosophy and science were nearly indistinguishable. The word “metaphysics” literally means “after physics” and I’ve heard philosophers joke that it’s called that because historians of old didn’t know what to else call the book of Aristotle’s writings that sat on the shelf next to “physics” on sort-of related topics.

Philosophy and science still share some common aims and methods, though philosophy is beholden to a somewhat different sort of “data” than the empirical sciences. (To be clear: we’re beholden to the empirical data too). Since we’re primarily interested in clarifying the concepts that map to the empirical data, a large part of our philosophical “data” is intuition and it is tested with things like “thought experiments.” We test concepts by asking whether they apply in certain kinds of hard theoretical cases and seeing how widely our intuitions are shared.

In some cases, thought experiments get turned into empirical studies. Practitioners of experimental philosophy (or ‘X-phi’) test whether intuitions are as widely shared as we like to think they are. If they find competing intuitions, they can challenge prevailing philosophical theories and provide new directions for theory.

All this is to say that the relationship between science and philosophy is complex and there are a couple of ways that I might apply my philosophical expertise as a data scientist. I could deploy philosophical concepts and frameworks in formulating questions and analyzing the data. Or I could do X-phi: test philosophical concepts by looking at data on intuitions and thought experiments.

Regardless of what I do, philosophy has given me a number of tools that should come in handy in my data scientist’s toolbox:

I’m well prepared to frame the question for my data’s audience. Why does any of it matter?

Also as a philosopher, my craft is logic and critical reasoning. These general skills are, of course, no substitute for area knowledge. Logically consistent nonsense is still nonsense. Nevertheless, it’s important to be able to identify gaps and inconsistencies in logic and make implicit premises explicit regardless of the topic. I can do that.

I hope you’ll come back and read more about my story as it unfolds in the coming weeks.

**Special thanks my Facebook friends who shared their thoughts on how to approach this topic when I was writing this.**