Twitch_City Twitch_City - 3 months ago 33
R Question

Incremental Plot in R Notebook Error

I am trying to create an incremental HE plot in an R notebook that will be compiled to HTML (so a .R file). The function I am using

heplot()
from the
heplots
package uses an
add = TRUE
parameter to overlay the graphic over the previous one, which is useful when there are multiple groups you wish to compare.

If I run this as an R notebook, I get the following error:

Error in polygon(E.ellipse, col = last(fill.col), border = last(col), :
plot.new has not been called yet
Calls: <Anonymous> ... withVisible -> eval -> eval -> heplot -> heplot.mlm -> polygon


I believe this is because the R notebook is not keeping the previous plot in memory when it evaluates the second plot.

Here is a reproducible example of the problematic R notebook file (save as .R):

#' ---
#' title: "Incremental HE Plots Test"
#' author: "Matthew Sigal"
#' date: "08 Jun 2016"
#' ---

#' ## Load package and data:
library(heplots)
data(Rohwer, package="heplots")

#' ## Multivariate models for two subsets:
rohwer.ses1 <- lm(cbind(SAT, PPVT, Raven) ~ n + s + ns + na + ss,
data = Rohwer, subset = SES == "Hi")
rohwer.ses2 <- lm(cbind(SAT, PPVT, Raven) ~ n + s + ns + na + ss,
data = Rohwer, subset = SES == "Lo")

#' ## Overlaid visualization:
heplot(rohwer.ses2, col = c("red", rep("black",5), "blue"),
hypotheses = list("B=0, Low SES" = c("n", "s", "ns", "na", "ss")),
level = 0.5, cex = 1.25,
fill = c(TRUE, FALSE), fill.alpha = 0.05,
xlim = c(-15, 110), ylim = c(40, 110),
label.pos = c(1, rep(NULL, 5), 1))

#' ## High SES students:
heplot(rohwer.ses1, col = c("red", rep("black", 5), "blue"),
hypotheses = list("B=0, High SES" = c("n", "s", "ns", "na", "ss")),
level = 0.5, cex = 1.25,
add = TRUE, # place both plots on same graphic
error = TRUE, # error ellipse is not drawn by default with add = TRUE
fill = c(TRUE, FALSE), fill.alpha = 0.05,
xlim = c(-15, 110), ylim = c(40, 110))


I thought that maybe using a chunk option, such as
fig.show="hold"
might work, but this did not solve the issue.

If I knit this in an Rmarkdown document, it works as expected (save as .Rmd):

---
title: "Rmd Test"
author: "Matthew Sigal"
date: "June 9, 2016"
output: html_document
---

## Test

```{r}
library(heplots)
data(Rohwer, package="heplots")
rohwer.ses1 <- lm(cbind(SAT, PPVT, Raven) ~ n + s + ns + na + ss,
data = Rohwer, subset = SES == "Hi")
rohwer.ses2 <- lm(cbind(SAT, PPVT, Raven) ~ n + s + ns + na + ss,
data = Rohwer, subset = SES == "Lo")

heplot(rohwer.ses2, col = c("red", rep("black",5), "blue"),
hypotheses = list("B=0, Low SES" = c("n", "s", "ns", "na", "ss")),
level = 0.5, cex = 1.25,
fill = c(TRUE, FALSE), fill.alpha = 0.05,
xlim = c(-15, 110), ylim = c(40, 110),
label.pos = c(1, rep(NULL, 5), 1))
heplot(rohwer.ses1, col = c("red", rep("black", 5), "blue"),
hypotheses = list("B=0, High SES" = c("n", "s", "ns", "na", "ss")),
level = 0.5, cex = 1.25,
add = TRUE, # place both plots on same graphic
error = TRUE, # error ellipse is not drawn by default with add = TRUE
fill = c(TRUE, FALSE), fill.alpha = 0.05,
xlim = c(-15, 110), ylim = c(40, 110))
```


So, my question is: how can I get the Rnotebook compiler to act similarly to the Rmarkdown compiler?

Answer

Apparently, my issue was very minor - my comment for the second HE plot created a new chunk, which is what caused the error.

A working script simply removes the #' from the comment between the two calls to heplot()!