Want to share your content on R-bloggers? click here if you have a blog, or here if you don’t.
Just some thoughts about the blog, R and whatever else I think I should learn..
The blog
I didn’t get much blogging done last year. I am at the point now where having older children means less time in the evenings to code, and I’m fine with that. Like many (I suspect) I have moments when I question the value of having a blog. I used to enjoy writing articles, then it began getting tedious. And it’s not the writing, it’s the technical upkeep, faffing around with images and links, making sure everything works, and then the blog architecture itself.
It seems just as you finally iron out all the background kinks, that’s when you get sick of the sight of it. So – a blog refresh is on the cards for this year. Managed to avoid blogdown, Hugo and distill.. and while it feels Quarto is inevitable, I can’t get excited about it. It’s too much like actual work. So it’ll probably be yet another jekyll theme.
I also need to think about the purpose. I don’t need to make the case for R in the NHS any more. Over the last few years its evolved into how I used R to tackle problems that have cropped up during work, so I suppose I should just continue down that road a bit more. One thing I do need to get comfortable doing is quick, “today I learned” type posts, because I have too many gists and other bits lying around with small examples of things I’ve used to solve various problems.
Newsletter
Does anyone need another newsletter? If I’ve learned anything, it’s that no one really reads blogs any more, so if nothing else, I could use this as a prompt to push people to check out my latest posts. I will have to find another host for it, because I am almost at a chargeable number of recipients and it makes no sense to pay for it as things stand.
R – what else is there to learn?
I have two things I’d like to improve on. My base R skills, and my working with deeply nested lists. I avoid the latter as much as I can, because they are horrible. Much of this is down to not having the experience of getting myself back out of the mess, so I need to to work on that. I’m also not sure how much more of my time I want to devote to R outside of work, because it feels it is very much a case of diminishing returns for me career wise. I’m proficient enough with tidyverse and data.table. I have packages on CRAN. I know how to use ggplot2, and Rmarkdown, and officedown, and targets, and duckdb. That’s plenty enough to deal with most of the tasks I’ll ever need to complete.
Excel
No, not kidding. There’s no denying that Excel has gained a lot of new functionality in the last couple of years (which I put to good use, developing a runcharter-esque spreadsheet that captured up to 5 new runs “automatically”). I know there are some NHS analysts who’ve never used Excel, but it’s the bread and butter (or cockroach, depending on your point of view) of analytics and I’m simply not comfortable having gaps in my knowledge, when I used to be seen as an expert (before I discovered R). I know exactly how to address these gaps, and it won’t take long, I just need the time and motivation. Imagine if I fell in love with Excel again and chucked R in the bin? Absolute scenes.
DAX
There is no escape from Power BI. Anyone can knock together a dashboard using it. Creating a well designed one takes some more effort, and at some point I need to immerse myself in DAX. I have been able to develop a suite of SPC charts using DAX ( e.g. laney corrected U-charts, xbar-s charts) and they have been very complex. At no point did I feel like I could use prior knowledge from similar tasks to arrive at a solution, it was a lot of trial and error and hoping that the next google result was better than the previous one. I’d like to firm up my understanding of what the different filter contexts were doing. Again, it feels like if you’re an NHS analyst, you’re expected to know PowerBI and Excel. Imagine if I fell in love with Excel AND PowerBI and….nah.. can’t see it. I think Power Query is great. DAX seems fairly incomprehensible a lot of the time, but often that is down to most examples being business/corporate data and metrics which are hard to map to healthcare.
Python
I have been coming to the conclusion that I need to add Python to the skillset. I’ve decided on the the first course I’m going to try and follow – got to the first part where I have to write some actual code, and anaconda fell over completely. Have had to uninstall/ reinstall and then immediately update software. By this time the children are needing assistance with other things and so the fleeting moment of opportunity is lost. Not ideal. Anyway, I shall try and persevere, just to see if I can gain any sort of excitement about it. Imagine if I find out zero-based indexing is what I’ve been missing all along? Yeah, not likely is it? Still, it’s everywhere, and I need to get on with it.
SQL
I should add Postgres or duckdb, or some other tool to the list, to add to MSSQL, although this is a much less pressing need.
So, apart from refreshing the blog, catching up on Excel, understanding DAX and finally learning Python, not much planned for the year.
We all know this is not going to happen, but, well, we can aim for the moon and land among the stars eh? And hopefully this time next year we can gather together as I recount the myriad ways of how I failed to achieve any of it.
Happy New Year!
Related
