…or, How I learned to stop working so hard and pad out my resume
I’m married. Like many married people, a honey do list is a fact of my life. My wife is a biochemist, so sometimes in addition to fixing things around the house my honey do list includes data analysis and/or cleanup on CSV files. I was a sysadmin in a past life, so for the cleanup side of things I turn to sed and awk first (that linked book is awesome). I’m no stranger to python’s csv module for those times when I’ve had to get to the next level with it.
With the actual analysis, I haven’t gotten around to learning numpy and/or scipy, but that linked book is in my queue. I have messed with R a bit and like it a lot for, well, messing around. It’s not hard to import a CSV or tab-delimited file; there are plenty of good “out of the box” functions; and I’ve managed to muddle my way through clustering and heat maps without too much swearing. So for doing a few analyses by hand or cranking out some quick graphics, it’s pretty handy for cutting to the chase.
Having cut my teeth a little, I wanted to go back and formalize my R skillz so I could put it on my resume (Hey! —Ed.) without fear of looking too ignorant when being called out. Different people have different learning styles in different contexts. (In yet another past life I was all up in some Math Ed stuff, but I’ll try to go light on the educational psychological jargon.)
As a “busy web professional,” I often learn things in a “just in time” manner. To get it to actually stick, I like to work on a toy project — I can often get more out of an hour of homework than a few hours of lectures. I’m good at closing out tickets and delivering deliverables, but when confronted by interview questions like, “Did you take course X while majoring in CS and do you remember topic Y?” my answer is inevitably, “No, I studied social sciences and no one cared what I majored in when I got my first computer job during the dotcom boom.”
Seriously, though, there are things you are “supposed” to know, especially if you say you “know” them on your resume (HEY! —Ed.). So I started looking for a way to be pseudo-credentialed in R where I could work on those credentials by punching the keyboard instead of punching my monitor.
My partial solution to this has been signing up for Coursera classes. I did this because there are a few different stat and/or R and/or biology data analysis-related classes on there. This is nice because there is homework, which caters to my learning style. The thing about Coursera classes that I find monitor-punchy is the lectures. Sometimes I will skim the lecture note PDFs or, under duress, crank the lecture playback up to 2x and then jump into the homework. Sadly, the lecture notes are often just that — a crutch for whatever they’re actually saying in the video, which I may or may not watch (usually may not because I have my real job and real life to worry about, too). Plus, this style of finishing the class to get a certificate risks sliding down the slippery slope into the purely “just in time” understanding that I was hoping to avoid, where I finish my homework… just in time.
Fortunately, I have found a good source of supplementary reading: the website that I conveniently work on. It’s faster and a lot less annoying to look up some details in R in a Nutshell than to try to figure out which part of a 30 minute lecture to sit through. Similarly, I find that when a book addresses the topic I’m trying to learn about in depth, it has more context than a
man page or an old email thread about someone else’s (hopefully) related problem. Even in the much-hyped age of digital learning, it’s still convenient to be able to skim through a section or two, get to the meat of what I’m messing with, and then go back and hack a little on my homework or side project. I’ll usually do this as a little bedtime reading; I find reading more relaxing than watching something before going to sleep.
D3 and Backbone.js
But wait, there’s more. I’ve been wanting to pad out my resume (!!! – Ed.) with some new hotness or with some other things that Coursera doesn’t have classes on. Continuing with the stat/visualization theme, I’m reading through Getting Started with D3. Tech books always have “toy programs,” but personally they don’t always satisfy, but having done some other homework or moonlighting in data analysis I can scrape together enough grist for the mill.
The book provides enough D3 to be dangerous after a few bedtime reading sessions, in a presentation that’s a little more hand-holdy than API documentation (although in D3’s defense they do have a pretty nice listing of tutorials). This backbone.js book is a better example of something I’ve found much more helpful that what I’ve seen online. So I already have enough in hand to pad out my resume a little more. (Sigh. – Ed.) Now if I can only find the spare time to hack on it…