Changing geom_point shape for specific values in a line - r

Question

I have a line graph with several lines, in which I want to highlight certain values with a different geom_point shape.
The dput of a similar data to mine is:
structure(list(Iso = structure(1:9, .Label = c("a", "b", "c",
"d", "e", "f", "g", "h", "i"), class = "factor"), z1 = c(342.6,
8.94, 6.91, 3.96, 1.89, 4.38, 1.43, 5.18, 189.1), z2 = c(187.34,
2.8, 8.42, 8.24, 2.36, 2.34, 7.6, 0.5, 136.01)), row.names = c(NA,
-9L), spec = structure(list(cols = list(Iso = structure(list(), class = c("collector_character",
"collector")), z1 = structure(list(), class = c("collector_double",
"collector")), z2 = structure(list(), class = c("collector_double",
"collector"))), default = structure(list(), class = c("collector_guess",
"collector")), skip = 1), class = "col_spec"), class = c("spec_tbl_df",
"tbl_df", "tbl", "data.frame"))
This is the code:
library(readr)
library(ggplot2)
library(RColorBrewer)
y <- read_csv("dummy.csv")
y$Iso <- factor(y$Iso, levels=y$Iso)
##
plot <- ggplot(y,aes(Iso,group=1)) +theme_bw() + ggtitle('') +
theme(plot.title = element_text(hjust = 0.5)) +
geom_point(aes(y=z1, colour='z1'), na.rm=FALSE, size=3, shape=16) +
geom_point(aes(y=z2, colour='z2'), na.rm=FALSE, size=3,shape=16) +
geom_line(aes(y=z1, colour='z1'), na.rm=FALSE, linetype=1,size=1) +
geom_line(aes(y=z2, colour='z2'), na.rm=FALSE,linetype=3, size=2) +
xlab('') +ylab('ZZ/CI') + scale_y_log10(limits=c(0.1,2000), breaks=c(1e-1,1,1e1,1e2,1e3))
##
plot + theme(axis.text.x = element_text(angle = 90, vjust = 0.5,size=19, face='bold'),
axis.text.y = element_text(size=12),
axis.title.y = element_text(size=15)) +
scale_colour_manual(name='',breaks=c('z1','z2'),values= c('brown','wheat'),labels= c('z1','z2')) +
guides(colour = guide_legend(override.aes = list (size = c(.75,.75),linetype=c(1,3),shape=c(16,16))))
I use the guide for every cosmetic element in order to make the legend more legible.
The data
# A tibble: 9 x 3
Iso z1 z2
<fct> <dbl> <dbl>
1 a 343.6 187.34
2 b 8.94 2.8
3 c 6.91 8.42
4 d 3.96 8.24
5 e 1.89 2.36
6 f 4.38 2.34
7 g 1.43 7.6
8 h 5.18 0.5
9 i 189.1 136.01
For example, in the case of these data, I want to change the shape from 16 to 13, in z1 for '342.6','1.43', and '5.18', and in z2 for '187.34,'2.36','7.6' and '0.5'.
How can I proceed?
Thank you for your time.

Answer 1

I think you should consider reformatting your data to long format, as it saves a lot of duplication. Secondly, you need to specify a new column in your data, which you map to shape. If there is no pattern like any value larger than 100 you need to cherry-pick the values you want to highlight by hand.
The following code should do what you want:
library(tidyverse)
y <- structure(list(Iso = structure(1:9, .Label = c("a", "b", "c",
"d", "e", "f", "g", "h", "i"),
class = "factor"),
z1 = c(342.6, 8.94, 6.91, 3.96, 1.89, 4.38, 1.43, 5.18, 189.1),
z2 = c(187.34, 2.8, 8.42, 8.24, 2.36, 2.34, 7.6, 0.5, 136.01)),
row.names = c(NA, -9L),
class = c("data.frame"))
## transform to long format
y.long <- y %>% gather(type, value, -Iso)
## add a new column which 'marks' the special rows
## NOTE: since we moved to long format rows corresponding to z2 are starting now at row 10
y.long <- y.long %>% mutate(highlight = ifelse(type == "z1",
ifelse(row_number() %in% c(1, 7, 8),
"special", "normal"),
ifelse(row_number() %in% c(10, 14, 16, 17),
"special", "normal")))
## in your ggplot you can now map the columns to the graphical elements like so:
ggplot(y.long, aes(Iso, value,
color = type, linetype = type, shape = highlight,
group = type)) +
geom_point(size = 3) +
geom_line(aes(size = type)) +
scale_y_log10(limits = c(0.1, 2000), breaks = c(1e-1, 1, 1e1, 1e2, 1e3)) +
scale_color_manual("", values = c(z1 = "brown", z2 = "wheat")) +
scale_size_manual("", values = c(z1 = 1, z2 = 2), guide = "none") +
scale_shape_manual("", values = c(normal = 16, special = 13), guide = "none") +
scale_linetype_manual("", values = c(z1 = "solid", z2 = "dotted"), guide = "none") +
labs(x = "", y = "ZZ/CI") +
theme_bw() +
theme(axis.text.x = element_text(angle = 90, vjust = 0.5, size = 19, face = "bold"),
axis.text.y = element_text(size = 12),
axis.title.y = element_text(size = 15))
This results in the following plot:
You can adapt the mutate satement to include/exclude other rows and the scale_* functions to show/hide legends.
Theoretically, you could use something like this in your mutate
mutate(highlight = ifelse(value %in% c(343.6, 1.43, 5.18, 187.34,
2.36, 7.6, 0.5),
"special", "normal"))
but due to floating point issues (cf. for instance this article), i would not do this and rather select by row number (as I did) or any other suitable criterion, or use all.equal to make proper floating point comparisons.