Things I liked this month – August 2017

I currently work on a team of software engineers. Some of them get scandalized by my workflow, processes, how my git repos are organized, etc., largely because (I think) there’s an implicit assumption that traditional software development practices apply to any field involving lots of code. I thought this article was a really helpful way to think about the relationship between data science and software engineering, especially insofar as it explains why data scientists are right to depart from many of the norms engineers are taught to follow.

Having a background in Philosophy from a Philosophy department that, strangely, produces as many data scientists as it does Philosophy professors affords me a strong interest in Philosophy of Data Science. My research was a graphical model of definitions, and so I’m super interested in defining data science. I’m writing on it. Slowly. Someone else is taking the question seriously, too, and I’ve found his writing really insightful.

Looker has made a data exploration VR. Cool AF.

Many people who know me also know that my opinion of data science recruiters as a group is very low. The overwhelming majority lack even the most basic understanding of what data science is or what skills are relevant. I’ve found one exception, who also wrote the best article I’ve ever read on how to be come a data scientist.

This. So much this. One of my greatest professional frustrations has been businesses that don’t understand that data science isn’t primarily about deploying software (or worse, that data science is little more than making pivot tables at scale), but science. Real science, in the same sense that we would say a theoretical physicist or chemist is doing science. Deploying software is crucial for turning science into dollars, to be sure, but the need for engineering is secondary; its value is entirely dependent on the rigorous scientific research it implements. In fact, my opinion that a physicist’s skillset is the most valuable one for data science, even more so than the skillset of a computer scientist or pure statistician, only continues to strengthen over time.

Leave a Reply