I've been updating some of the lectures for the health data science lectures & courses I teach, and I wanted to take a minute to reflect on how much evolution it feels has happened in the last 5-ish years.
When I started teaching and advising students on data science years ago, I introduced my perspective on ds (mainly to epidemiology and biostats students) as contrasted to other, narrower fields using this 3-axis model. My perspective was that working in ds in healthcare meant some combination of these 3 areas.
Compare my previous 3-axis model to this radar chart I've updated into in an attempt try to be more comprehensive of the disciplines commonly coming up in modern data science work. Based on conversations with other data scientists and employers I've identified 8 high-level skills / disciplines that can comprise ds work and many of those are aggregated (i.e. data engineering represents aspects of architecture, warehousing/databases, data integration).
Data science has become a VERY BIG tent, which is great since we represent so many different backgrounds, training, and skillsets. But what data science means for many of us day-to-day and especially talking with employers feels like it's gotten more ambiguous, not less. I've been advocating that data science move to something like my radar-chart based 'data science thumbprint' I teach to quickly understand what 'data science' means to an individual or company. How great would it be to see something like this in a job posting as opposed the common generic "strong experience with R, SAS, or python & some experience writing JAVA, C++, or Scala"?
