I am a beginner of xyplot in lattice package, and I am trying to visualize my data frame. I would like to present values of some continuous variable on x-axis and a couple of factor variables on y axis (In other word, nested variable on y-axis)
Q: How can I add custom strips labels to y-axis?
I read the manual document and googled this question while no solution was found. What I expect to obtain is something like this
Is that possible to get this kind of plot using xyplot?
Update
Indeed, ggplot2 might be a good option. However, xyplot seems to be more convenient, through formula as y~x1|x2, while I do not know how to make that plot.
Here is a reproducible example
method<-rep(c("XGBM", "LGBM", "WGBM", "NNGBM"), times = 20)
Scenario<-as.factor(sort(rep(1:5, times = 16)))
scrooter<-as.factor(rep(c(50, 5, 0.5, 0.05), times = 5, each = 4))
measure<-rnorm(80)
data<-data.frame(measure, method, Scenario, scrooter)
str(data)
'data.frame': 80 obs. of 4 variables:
$ measure : num -1.181 -2.595 1.114 0.178 -1.246 ...
$ method : Factor w/ 4 levels "LGBM","NNGBM",..: 4 1 3 2 4 1 3 2 4 1 ...
$ Scenario: Factor w/ 5 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
$ scrooter: Factor w/ 4 levels "0.05","0.5","5",..: 4 4 4 4 3 3 3 3 2 2 ...
measure: circles in the plot and values should be placed on x axis.
method: strip labels on the top
Scenario: strip labels on the left
scrooter: the value placed on y axis
Update 2
Here is a solution using ggplot but I am looking for the one using xyplot
cust_theme <- theme_bw() +
theme(legend.position="none",
panel.spacing.x = unit(0, "lines"),
panel.spacing.y = unit(0, "lines"),
panel.grid = element_blank(),
panel.border = element_rect(size = 0.25, color = "black"),
strip.background.x = element_rect(fill = "wheat1"),
strip.background.y = element_rect(fill = "lightgreen"),
strip.text.x = element_text(margin = margin(1,0,1,0, "mm")),
strip.text.y = element_text(margin = margin(0,1,0,1, "mm")),
axis.title=element_text(size=14,face="plain"),
axis.text.x = element_text(angle = 45, hjust = 1)
)
point.plot<-ggplot(data = data, aes(x = measure, y = scrooter)) +
geom_point(size = 3) +
facet_grid(Scenario ~ method, switch = "y") +
cust_theme +
geom_vline(xintercept = 0, linetype="dotted", color = "black", size=0.5) +
ylab("Scrooter") +
geom_point(color = "white", size = 1.5)
The key function is useOuterStrips()
available in the latticeExtra
package. It works on a trellis object with exactly two conditioning variables, as you have here. You can wrap it around a function call as shown in the first example or, as is more typical, apply the function to the saved the trellis object as shown in the second example.
# Starting with data as specified above
# Need latticeExtra, use two steps to build plot
require(latticeExtra)
# Wrapped around a call to xyplot with two conditioning factors
useOuterStrips(xyplot(scrooter ~ measure | method + Scenario, data,
panel = function(...) {panel.refline(v = 0); panel.xyplot(...)}))
# More typically, and to more closely reproduce the example here
obj <- xyplot(scrooter ~ measure | method + Scenario, data,
panel = function(...) {panel.refline(v = 0); panel.xyplot(...)},
scales = list(alternating = FALSE, tck = c(0.5, 0), col = "gray40",
x = list(rot = 45)),
as.table = TRUE, ylab = "Scrooter", col = 1)
# Apply useOuterStrips function
useOuterStrips(obj)