If you’re diabetic, blood glucose testing is something you live with daily, but it’s good for the rest of us to test from time to time as well. That’s what I did this morning using a Contour Next EZ blood glucose monitor, which at $10 (or free with a doctor’s prescription and a purchase of 100 strips) is the cheapest device I know.
Serious testers will want something that does continuous monitoring and connects to your computer for easy download, but I just wanted something simple and cheap. If you can get a doctor to prescribe it, I’ve heard the Dexcom CGM device is the way to go, but this is a simple and cheap way to get started and I just want to demonstrate briefly how easy it is to generate a glucose level graph in R.
To make it interesting, I tested myself immediately after waking up and then drank a glass of sugar water. After waiting another half hour, I tested again, then again at two more intervals after that. I tracked my blood glucose levels and then wrote the following R script.
First, load the prerequisite libraries. It’s super-easy if you use Tidyverse. In this simple example I won’t use any of the Tidyverse features other than graphics, so all of this extra loading is unnecessary, but I think it’s a good idea to get used to treating Tidyverse as an extension of your normal R environment. It doesn’t take much to load it, so why not? Just ignore all the “Loading” and “conflicts” messages.
## ── Attaching packages ────────────────── tidyverse 1.2.1 ──
## ✔ tibble 2.1.1 ✔ purrr 0.3.2 ## ✔ readr 1.3.1 ✔ forcats 0.4.0
## ── Conflicts ───────────────────── tidyverse_conflicts() ── ## ✖ stats::filter() masks dplyr::filter() ## ✖ stats::lag() masks dplyr::lag()
Next I’ll set up my data. These are actual numbers taken from the glucose monitor. As you can see, I’m creating a simple data frame in R code. If this were a full example, I’d have the data in Excel or as a CSV file and then load it using
read_excel(), but I don’t want to deal with that now, so I’ll construct the data frame by hand:
glucose <- c(85,141,75,74) times <- c(0,30,30+45,30+45+20) measures <- data.frame(glucose=glucose,time=times) measures
## glucose time ## 1 85 0 ## 2 141 30 ## 3 75 75 ## 4 74 95
To plot it, I can simply run the
ggplot function below, which will generate the default plot style or “theme”. But to make it look a little more cool, I can use the
ggthemes library, which gives you a choice of themes, everything from “fivethirtyeight” to “tufte” and “wsj”. Here I’ll make my final result look like one of those charts you see in The Economist newsmagazine.
library(ggthemes) theme_set(theme_economist()) ggplot(data=measures,aes(x=time,y=glucose)) + geom_line() + ggtitle("My Glucose Levels") + xlab("Time (min after waking)") + ylab("Glucose (mg/dL)") + annotate("text",x=10, y = 68, label = "Drink Sugar Water", color = "red") + annotate("segment", x = 00, xend = 0, y = 70, yend = 80, colour = "red", size=1, arrow=arrow())
That’s it. Although this was a very simple – some would say trivial – example of graphing in R, I hope it’s helpful for absolute beginners who want to know how to get started.