Member-only story
Boot Camp Tales
A Curriculum Synopsis
Somewhere in one of our bootcamp slide decks we define a “data scientist” as someone who knows more about programming than your average statistician, and more about statistics than your average programmer. I found that clever.
Note that this definition doesn’t preclude knowing more about programming than your average programmer, nor being more stats savvy than your average statistician; these extremes could be included, but would be a measure of strength over and above what we call setting a lower bar.
As a code school instructor, my job is to look towards raising my students above that lower bar, and at the end of 2024 this means taking the Python path into programming, and a machine learning (ML) approach into stats.
I’m like the conductor slash tour guide from London to Berlin, on a train bound for Beijing eventually. We start with basic Python and work our way up through Data Analysis and then Data Visualization. These three are separate modules, with more modules to follow.
When working through Python basics, I start mentioning statistics in an historical context, talking about:
(a) how in my experience the stats folks have always been hungry for computer power and
(b) how we might explore the subject by following some of it fissures and controversies.
That’s one of my chief learning techniques: discover the front line, most heated debates, including past ones they may have since cooled, to get the flavor of what’s in store. To get a sense of what’s ahead, study what’s been happening. A truism perhaps, yet it deserves to be said.
Caveat: “stats folks” doesn’t mean only pro statisticians but includes those professions that use stats a lot, for example medical and astronomical, in addition to within certain disciplines, such as politics and economics.
Statistics is stereotypically a dry topic such that the notion of “heated debates” might strike some as oxymoronic, yet indeed such debates have occurred, around the very ontology behind the subject itself.
Of course I’m talking about the Frequentist objectivist versus Bayesian subjectivist tension, where the latter is coming out on top in machine learning…
