R Programming for the elderly …
As a teacher and future masters student, notwithstanding my imminent fifty-eighth birthday, I am interested in data science. Some recent post-graduate study required a small piece of original research which served as an appetiser both in terms of collecting data and then analysing it. I enjoyed the process and got a good grade for my submission. It’s self-evident, isn’t it, that any teacher will do a better job if they approach teaching with knowledge, backed up by research, about what actually works in the classroom ?! So over the last few months I’ve signed up for a mooc or two to keep my hand in pre-masters and as a significant element of said masters will be a 12–15,000 word research project, what better subject area to study than data science.
Which leads me to R and statistics. First of all I do not understand regression despite the very photogenic Dutch lecturer’s assertion, at the end of the third video lecture, that I do now I have watched his video. He speaks wonderful, flawless English too, but that doesn’t matter, I am still not getting it. As part of this mooc I have to do ‘data camps’, (think online concentration camp), in order to learn R: machine learning, (and I apologise to anyone who had the stupidity to read the comments I was invited to leave on popup screens after umpteen meaningless exercises), designed to get learners up to speed in this programming language. So far it is a humiliating waste of time!
On realising this wasn’t working like a good student I sought inspiration from the mooc discussion board. From a quick perusal of all these comments which indicated my fellow students were devouring this course with lusty appetites, and apparently all were advanced programmers in less than a week, I decided to sign up for additional moocs. I am happy to say these new courses have helped me to learn one useful fact, utterly unrelated, however, to learning R. It is this: it’s called Johns Hopkins University and not John Hopkins University. I never knew until signing up for these courses that it was Johns with a final ‘s’ and not John. Now that is really interesting and I am pleased to have learned this.
One of the JHU moocs has another machine learning platform called ‘swirl’, a rather unfortunate mnemonic, methinks, due to its strong porcine associations. Anything to do with pigs reminds me of Sean who used to say in his drole Edinburgh accent when describing his misspent yoof: “ah moved oot an’ the pigs moved in”, ( ‘When I moved out, the pigs moved in!”). I think ‘swirl’ in this respect has no relationship with ‘swill’ but rather stands for ‘statistics with interactive R learning’. The first week’s course content lists seven ‘swirl’ sessions with an estimated time of three hours for each. Ouch! That’s twenty-one hours calculated by old-fashioned mental arithmetic, no machines necessary. Fortunately the sessions are voluntary — phew! The good news, and perhaps proof of the existence of a loving God, is I cannot get ‘swirl’ to open in my R console, so I don’t need to find any hours at this moment to complete this jolly jape.
However I felt really good about being able to download swirl and then check its version by using the command >packageVersion[“swirl”]. I even managed to download the file for the Quiz Data Set, copy it into my working directing , (WD!), and set my working directory to the folder that includes this file, but I cannot get R to read it for me. No problems really because I can complete 90% of the quiz by extracting the data from the original Excel spreadsheet. Right, yeah, I’m supposed to be doing this with R — sorry, no can do, can’t get the programme to read the csv. file and I need to get that 100% grade and meet the deadline. Good old Excel, huh, never lets you down, possibly the best bit of software ever designed by Microsoft!
Finally an absolute must for all anoraks: how to ask questions the smart way from which I learned the meaning of ‘RTFM’ and ‘STFW’. Let me help you with the middle two letters : …….. the f****** …….. and ……….the f****** ……….. ! I posted a plea for help in a discussion board on how to use swirl on R, based on recommendations in this fine document, and 24 hours later, further proof of that loving God can be inferred from the total response of zero. If I had been suicidal, I would have jumped by now!
So I’m having a blast but I am not sure whether I am actually learning anything. Just as well I am on holiday. But I do want to learn this stuff so I am persevering and I promise to update you on progress.