On trying something new
As a grad student the work/life balance tends to blur. I actually do most of my work on the weekends and tend to try to take time during the week for myself. This works best for me because weekends are typically when no one needs me to physically be somewhere, so I can get into the correct headspace to do some work. This weekend was a busy one, and my Co-PI had some ideas about one of the papers I’m working on that required my full attention. He also wants me to do something I’ve never done before, so let’s talk about trying something new!
I love research because you’re constantly learning something new. It’s never stagnant and just when you get comfortable things shift. We do new experiments, look at things from different ways, try new techniques, or in my case try different analyses. I thought this weekend would be a slow couple of days. I had most of my work done, I was coming off of a super successful lab presentation, and I had only minor things to worry about with some of the papers I’m working on. Ironically one paper I’m working on I didn’t even mention recently because I’m “just” second author on it, so it never occured to me to include it in the long list of first author papers I’ve been working on.
Instead, I got an email from my PI which required me, among other things, to redo the analysis I had done on the dataset for our paper (which I am first author on, yay). It’s not the end of the world and I’ve talked about how this wouldn’t be the first time going back and redoing things (like this example). It did mean some extensive work on recreating all the figures the change in the analysis would alter, but in the grand scheme of things, that’s not what was going to take forever.
What was going to take forever was the request to do something new. I’ve never done it before and as usual I can’t talk exactly about what “it” is, but I can say that “it” will result in some data that I need to turn into a figure and that’s where I’m stuck. I need to figure out the best way to display the data and it will have to be some incredibly impressive way because my Co-PI isn’t super thrilled with my performance.
He didn’t say it quite like that, he said he felt like I was only giving it like 93% effort and not 100% effort. To which I responded he’s lucky his birthday has come and gone already. But in all seriousness I’ve put a lot of effort into this project and it kind of hurt to hear that, but I get why he did it. He wants me to do something better, I’m just not sure what better looks like at the moment, so I’m playing around with different ideas for how to display the data and make it look impressive. It takes a lot time time, tinkering, and editing code to make things look just the way you want them, so that’s where the bulk of my time will be spent.
PLUS, I need to first settle on a type of plot to use, which I have so many to choose from it’s hard to narrow it down because the data could be displayed several different ways and the best way to show it may not be the first way I choose. I’m going to have to play around with different styles to see which kind of graph I like best. Then there’s the software I want to use to make the plot, which I have two main options (the third would be excel, but come on, no one does that).
The first would be to use MATLAB. This would be my preferred choice of software to use in most cases. MATLAB is my “home” language and it’s what I know best. I can make figures and edit just about everything there is on the figure to make it perfect. Or at least my idea of perfect. MATLAB also makes some very nice figures (in my opinion) and is how I do most of my research figures these days. I mean I taught a weeks long MATLAB course over the summer (here) so it’s safe to say that I’m pretty skilled and that makes life easier because I don’t have to hunt down commands or search vaguely for the function I want to use.
The second option is software I’ve only recently (early this year) started using. Rstudio has been surprisingly useful in making figures and while the language itself is weird, to me anyway, it can make several different styles of plots easier than I could do in MATLAB. Basically ggplot2 has been a lifesaver in terms of generating some of the figures that I can’t do easily in MATLAB (if at all). Plus they have a handy website with a full assortment of figure ideas and example codes to help me select which style I want to use (here). If you’re using Rstudio for anything, I HIGHLY recommend that website.
The catch with the second option is that I’m not great at using the language for Rstudio. One of two things is happening, either I don’t exactly understand how to do something, or the packages (ggplot2) aren’t as flexible as what I’m used to using in MATLAB. I can’t easily change different parts of the figures around like I would do in MATLAB and some cases I can only use whole numbers when I want to use fractions, things of that nature. There’s probably ways to work around it, but I don’t know them! Changing fonts around also feels like a pain in certain cases, I don’t know, it’s just a lot of things. It’s still vaguely unfamiliar to me.
While I prefer MATLAB, I’ve been using Rstudio a lot for this paper. Mostly because of the stuff I just mentioned about certain types of figures being easier to make. In either case I’ve got a lot of work ahead of me and not a lot of time to do it. So I need to figure out first which plot style I like before I can settle on if I want to work in MATLAB or Rstudio, it’s probably going to come down to the style of figure, or maybe I’ll just be lazy and keep using Rstudio since that’s what I’ve used for most of this paper. It’s a tough choice and I won’t know if I’ve made the right decision until well after the fact.
The outcome (or lack of making a choice) really boils down to fear. The fear of trying something new and failing. No one wants to put all that effort into something only to be told it’s awful. While I hope that won’t be the case, it’s still something I’m scared of doing. I want to make something super impressive, but the only way that will happen is if I make something to begin with, so good talk I guess.
I’ve got some work to do and figures to make, but hopefully in a few months (less than a year I would hope!) I can share them here and you’ll all be impressed with how cool they look. Or it could be a total flop, but I won’t know until I try.