Ridiculous week of working in R, in RevBayes, and learning some new or at least newish modeling techniques in order to put together an integral projection model, which is way of understanding the vital rates of a population (growth, survival, reproduction) and making predictions about the future. Very dense statistical week. I feel like I’ve been in the clouds every day, all day long, and then sort of having stress dreams in various programming or math languages all night.
A pic of a really interesting cat the Sig O was fostering for a while, who was probably developmentally divergent and had been through a house fire. And a fig.
I finally realized today that I wanted to deploy the exact same method, pretty much, that a group of researchers from Mexico had already used and published, and I knew three of them, so I contacted them and proposed we meet and see if it’s a good fit for collaboration, and discuss co-authorship if so. One of them already has all the R code I need, and that is pretty amazing, considering that realistically it would take me a few months of experimentation to put it all together. So I hope they agree that it’s a good collaboration and I can just start using the existing code.
The above sweet mass produced cakes are just awful, but when you’ve been camping on a remote island in the Pacific Ocean off the coast of Baja for two weeks, they are the very height of luxury.
Anyway, massively random post. Just want to check in fairly regularly even if I can’t describe exactly what is going on. What is going on? Tons of brain exercise and very little else, at least lately. It is remarkable how I sometimes just disappear for hours into a whole universe of learning new things. Today, I was doing an online tutorial of how to build a generalized additive model in R. The world of data analysis, statistics, and R is almost 100% online these days, so as soon as one learns how to follow along in one of these R tutorials, basically the sky’s the limit. It’s interesting that many modeling and other processes in R can be best learned by doing, even when the concepts themselves are incredibly dense. Over time, through iteratively doing the modeling and analyzing the results, the concepts start to make sense. It’s a lesson for me to remember that learning by hands on iteration is just as valid as having the entire conceptual framework down and then applying it.
My sense of how long some of these learning processes “should” take has definitely changed over the past five years or so. I used to get really down on myself for being “slow,” and for certain things that seemed like they should be the essence of simplicity to take me forever. Some of my younger colleagues had a lot of experience in coding and using R and they could just speed right through things that would sometimes take me hours. But now I know better, and whenever I am embarking on a new thing, even if it sounds simple, I am a ton more patient than I used to be. An example might be that you want to plot the standardized precipitation evapotranspiration index for Tucson from 1905 to the present. So you download the data, learn how to extract it. That process in itself took me a couple days. Then you get it in useable form, and you plot it.
Nothing wrong with this pretty stark black and white plot, very clearly showing times of wetter than usual (greater than 0) and drier than usual (less than 0) values.
But you’re looking at the plain plot and you start thinking about ways you could make it more informative. You get that dreaded feeling that you’re about to go down a ggplot2 rabbit hole. What if the values great than zero were green and the values less than zero were red?
Okay, that is kind of cool….but WHAT IF the red and green were colored on a gradient of values, with darker colors being higher absolute values and values closer to zero being more pale????? That process took me about a day, lol.
I have to admit I think the “final” version is better, and it was worth the extra time. A real wizard at R would have known right away how to get from the black and white one to the final one. It takes me forever, and I’ve been working in R since fall, 2014.
Anyway, as the weekend approaches, and as I relish the prospect of collaborating with a couple of other researchers who actually know what they are doing, I think maybe I’ll give myself a mental break.
Leave a Reply