Saturday, September 23, 2023

Thoughts on the Subtly Anti-Establishment Vibe of a Statistics Textbook




In the last decade scientists began using computer programs called Jupyter Notebooks, which combine formatted text with interactive graphs and executable snippets of code. Unlike a pdf or textbook, a Jupyter notebook allows a reader to do experiments within a text, modifying or adding mathematical code to assist learning and check comprehension. Notebooks also increase the potential for collaboration; a notebook file can be uploaded to GitHub (also a free resource) and revised by groups across a tracking history. These tools of knowledge transfer and retention--used across STEM disciplines and elsewhere--are vastly more powerful and affordable than the licensed tools available 10-15 years ago. If this does not seem remarkable, it's probably due to the corrosive effect capitalism has had on self-publishing technology.

As some may recall, it was once edgy and fun to maintain a blog. It did not seem to matter that blogging platforms had poor text formatting features and lacked computational or revision tracking features; some of us were thrilled to see our verbal ideas simply go public. At that moment in history one could assume the founders of Twitter realized that brevity would become a commodity in "cyberspace", a corrective market response to the masses writing to their hearts' content at different URLs.

Comments and likes have always measured blogging success, so Twitter shrewdly adopted those features. By about the mid 2010s, most bloggers had become "micro bloggers" on Twitter, and the ones who wanted to communicate at length ironically resorted to commenting on their own posts in order to work around Twitter's 140 character limit. (This practice is euphemistically called "posting a thread".) As the coup de grace to self-publishing's credibility, Elon Musk purchased Twitter in 2022, terminated most of its ethically-conscious staff, and smugly renamed the whole enterprise "X".

As Giles McMullen-Klein points out in a youtube video, even wonderful writing and scholarship tend to languish without the force of a well-funded marketing strategy. We see that the flip-side of viral cultural phenomena is the discounting of intellectual culture. But discovering a book like An Introduction to Statistical Learning with Python provided for free by Stanford U., and containing over a dozen Jupyter notebooks gives hope and encouragement. The fundamentals of machine learning and AI are basically all there, free for the taking and tinkering.

No comments: