Merry christmas, I guess?

I’m not a huge christmas guy. I have some small holiday traditions and my family tends to cook Thanksgiving part 2 on christmas day but otherwise I could take it or leave it. Like birthdays, as a child christmas is great but as an adult it becomes a mess of painful logistics, gut wrenching price tags, and excessive socialization.

I don’t dislike the holiday I just don’t think it’s a great holiday. It’s also the only holiday where actively voicing negativity towards it is met with public shame and concerns over childhood trauma. I grew up in a good, safe home. I learned of the falsehoods surrounding santa when I was 8. It wasn’t traumatic — if anything it made me feel better to know there wasn’t a magic man watching like a hawk for bad behavior so he could fill my stocking with mesothelioma. My psychiatric well-being and Freudian tendancies don’t need to be studied simply because I don’t find a Hallmark holiday celebrating Paegan indoctrination to be a sufficient reason to stress myself out traveling at the worst time of year only to give family and friends gifts I spent the year withholding from them. God forbid the gift is given or arrives after christmas, then I’m a selfish asshole.

Now what I do enjoy is setting goals. God, I love goal setting. Every game I play, I play because it involves setting and achieving goals. My career paths, my education, my hobbies; goal setting focused. New Year’s resolutions are about as silly in concept as christmas except they only (usually) affect me. Which means I can do dumb shit like set a goal that exclusively revolves around public self-flagellation. Oh look what a perfect segway to the public self-flagellation goal I’m setting this year!

Malcolm Gladwell is apparently a jackass? I guess don’t meet your heros (I don’t intend to for this one. What does a statistician even do with a journalist? Argue?) because I quite enjoy his writings and podcast. At the risk of throwing a couple punches at a sufficiently pulverized horse, I’d like to do a challenge loosely based on Gladwell’s 10,000 hours.

I’m exceedingly dyslexic (awful name for that condition by the way, violently abused vowels and silent letters for this kids who can’t spell?) but I’m enthralled by textbooks. Can’t stand fiction novels — if I’m going to torture myself I’d like to gain something out of it besides “a good story”. Textbooks are self paced college courses I can carry in a backpack which is fucking incredible. Since standard Goodreads “X number of books this year” challenges aren’t well suited for textbooks I need to attack a reading challenge from a different perspecitve. Meet: the 10,000 pages challenge.

Over the next year I’m going to be reading $\approx$ 30 pages a day from 25 different textbooks. Those 25 textbooks contain roughly 10,263 pages but if we consider that most of those textbooks have their page counts muddied up with indices, forwards, and random junk pages, that’s probably about 10,000. At that pace it’ll take me about 342 days to wrap up the challenge, allowing me a lot of wiggle room for cheat days and burn outs. I’ll be using this blog as a way to journal those 30 pages so that I (hopefully) learn something from them.

To indulge the morbid curiosity of onlookers, here are the books in no particular order:

• Code: The Hidden Language of Computer Hardware and Software by Charles Petzold

• Algorithms by Sanjoy Dasgupta

• The Algorithm Design Manual by Steven S. Skiena

• Concrete Mathematics by. Ronald Graham

• R for Data Science by Hadley Wickham

• Advanced R by Hadley Wickham

• Metaprogramming in R by. Thomas Mailund

• Functional Programming in R by Thomas Mailund

• Advanced Object-Oriented Programming in R by Thomas Mailund

• The R Inferno by Patrick Burns

• Tao Te Programming by Patrick Burns

• Refactoring by Martin Fowler

• Doman Driven Design by Eric Evans

• Proofs by Jay Cummings

• Real Analysis by H.L. Royden

• The Elements of Statistical Learning by Trevor Hastie and Rob Tibshirani

• Statistics for Experiments by George E.P. Box

• Statistics As Principled Argument by Robert P Abelson

• In All Likelihood by Yudi Pawitan

• Bayesian Data Analysis by Andrew Gelman

• Statistical Rethinking by Richard McElreath

• How to Lie with Statistics by Darrell Huff

• Fooled by Randomness by Nassim Nicholas Taleb

• Judgement Under Uncertainty by Daniel Kahneman

• The History of Statistics by Stigler

The R programming literature is rather incestuous since there aren’t that many good, passionate program R educators. For good reason — the language fucking blows. I just want to count myself as one of the best alive so that I can flex on Python brogrammers that their language may be powerful but it can’t fix their incompetence.

A lot of this list is meant to just nail the coffin shut on my skillset as a statistician but things like Royden’s Real Analysis are there for spite and spite alone. I’m not a very good mathematician and I never will be, but it’d be fucking hilarious if I was better at proofs than most “mathematical statisticians”. I know — it’s a low bar.

Otherwise that’s the deal. 30 pages a day, 342 days, plenty of room for error, no god damn shot I’m going to keep that pace with Royden but we ball.

I’ve done some cheating and read ahead on some of these and I’ve been consuming a lot of Burns’ “The R Inferno” while on break (because I cannot sit still). I don’t need R for Data Science in the same way an introductory user would need it but I should probably read it to say I have. Burns’ writing is a better representation of where I’m at in the rabbit hole currently.

41 pages deep and I gotta say, hilarious. Patrick’s writing style has the “edge” that I tend to carry around in my own professional communication. I’ve been warned by my “Ph.D. advisor” (the quotes are very necessary but hard to explain) that I need to soften the edges to survive in the world of academic statistics. His advice has been good so far but that one I’m going to classify as “cooked” and move along. If I can’t be publicly mentally ill then I may as well just go to industry and stay there. Being abrasive is worth the $>30\%$ paycut to me so the moment that’s gone I’m gone.

R Inferno is a great piece not just because of the writing style but the content itself. Burns isn’t just roasting R over the pits of hell, he’s advocating for more conscious usage of the language. A lot of R programmers are mono-lingual and unaware of proper programming priciples. I’m no saint here — I’m guilty of “speaking R with a thick C accent”. I’m better suited for lower level languagues because my mind solves problems algorithmically, something that isn’t intuitive to do in R.

So far I’d say the greatest value of those first 41 pages is Burns’ statements on the value of vectorization and abstraction. I’ve been spending a lot of my time trying to rationalize a divorce from R programming yet a bulky document comparing R to the circles of hell has thoroughly convinced me to make the relationship work. I’m operating from a Macbook with 16gb of RAM so I don’t get much say in the matter of efficient programming. I thought my release would come in the form of learning a more performant language but I’m starting to realize it comes in the form of being less shit at programming. Simple things like figuring out the mathematical “trick” to do a bulk of operations and finding places to replace loops with vectorized programs is going to make those 16gb feel less punishing. Already I’ve managed to replace the awful for loop I use in JAGs posterior prediction with a fully vectorized version by abusing the outer() function.

I’m not going to keep reading R Inferno much longer because I’d rather jump down In All Likelihood to try and reap some rewards of this challenge early, but I think I’ll snag another 30 pages tomorrow before I call it quits.

1
 41/10263 = 0.4% of the way there