Plot breaks due to deprecated function - r

I am trying to reproduce the plot from this question, but code is deprecated and I cant seem to figure out why it always gives the error.
Error: Discrete value supplied to continuous scale.
I thought I had omitted each row for being discrete to figure uot what was going on, but anyway I do it it alyways breaks because of that. There are some minor errors due to axis.ticks.margin and panel.margin as well as vjust but I don think they are the main issue. Although not 100%.
Find the reproducible dataset here:
groupData <- dput(structure(list(ID = 1:12, Group = c("Renal Failure", "Renal Failure",
"Diabetes", "Diabetes", "PA Disease", "PA Disease", "CV Disease",
"CV Disease", "Sex", "Sex", "Age", "Age"), Subgroup = c("No",
"Yes", "No", "Yes", "No", "Yes", "No", "Yes", "Female", "Male",
">70 yr", "<70 yr"), NoP = c(4594L, 66L, 2523L, 2228L, 4366L,
385L, 4296L, 456L, 908L, 3843L, 1935L, 2815L), P_S = c(0.2, 0.37,
0.84, 0.06, 0.37, 0.33, 0.18, 0.69, 0.21, 0.47, 0.17, 0.77),
P_G = c(0.51, 0.51, 0.13, 0.13, 0.54, 0.54, 0.41, 0.41, 0.46,
0.46, 0.46, 0.46)), class = "data.frame", row.names = c(NA, -12L)))
Code
## REQUIRED PACKAGES
require(grid)
require(ggplot2)
require(plyr)
############################################
### CUSTOMIZE APPEARANCE WITH THESE ####
############################################
blankRows<-2 # blank rows under boxplot
titleSize<-4
dataSize<-4
boxColor<-"pink"
############################################
############################################
## BASIC THEMES (SO TO PLOT BLANK GRID)
theme_grid <- theme(
axis.line = element_blank(),
axis.text.x = element_blank(),
axis.text.y = element_blank(),
axis.ticks = element_blank(),
axis.title.x = element_blank(),
axis.title.y = element_blank(),
axis.ticks.length = unit(0.0001, "mm"),
axis.ticks.margin = unit(c(0,0,0,0), "lines"),
legend.position = "none",
panel.background = element_rect(fill = "transparent"),
panel.border = element_blank(),
panel.grid.major = element_line(colour="grey"),
panel.grid.minor = element_line(colour="grey"),
panel.margin = unit(c(-0.1,-0.1,-0.1,-0.1), "mm"),
plot.margin = unit(c(5,0,5,0.01), "mm")
)
theme_bare <- theme_grid +
theme(
panel.grid.major = element_blank(),
panel.grid.minor = element_blank()
)
## LOAD GROUP DATA AND P values from csv file
groupData
## SYNTHESIZE SOME PLOT DATA - you can load csv instead
## EXPECTS 2 columns - integer for 'ID' matching groupdatacsv
## AND 'HR' Hazard Rate
hazardData<-expand.grid(ID=1:nrow(groupData),HR=1:6)
hazardData$HR<-1.3-runif(nrow(hazardData))*0.7
hazardData<-rbind(hazardData,ddply(groupData,.(Group),summarize,ID=max(ID)+0.1,HR=NA)[,2:3])
hazardData<-rbind(hazardData,data.frame(ID=c(0,-1:(-2-blankRows),max(groupData$ID)+1,max(groupData$ID)+2),HR=NA))
## Make the min/max mean labels
hrlabels<-ddply(hazardData[!is.na(hazardData$HR),],.(ID),summarize,lab=paste(round(mean(HR),2)," (",round(min(HR),2),"-",round(max(HR),2),")",sep=""))
## Points to plot on the log scale
scaledata<-data.frame(ID=0,HR=c(0.2,0.6,0.8,1.2,1.8))
## Pull out the Groups & P values
group_p<-ddply(groupData,.(Group),summarize,P=mean(P_G),y=max(ID)+0.1)
## identify the rows to be highlighted, and
## build a function to add the layers
hl_rows<-data.frame(ID=(1:floor(length(unique(hazardData$ID[which(hazardData$ID>0)]))/2))*2,col="lightgrey")
hl_rows$ID<-hl_rows$ID+blankRows+1
hl_rect<-function(col="white",alpha=0.5){
rectGrob( x = 0, y = 0, width = 1, height = 1, just = c("left","bottom"), gp=gpar(alpha=alpha, fill=col))
}
## DATA FOR TEXT LABELS
RtLabels<-data.frame(x=c(rep(length(unique(hazardData$ID))-0.2,times=3)),
y=c(0.6,6,10),
lab=c("Hazard Ratio\n(95% CI)","P Value","P Value for\nInteraction"))
LfLabels<-data.frame(x=c(rep(length(unique(hazardData$ID))-0.2,times=2)),
y=c(0.5,4),
lab=c("Subgroup","No. of\nPatients"))
LegendLabels<-data.frame(x=c(rep(1,times=2)),
y=c(0.5,1.8),
lab=c("Off-Pump CABG Better","On-Pump CABG Better"))
## BASIC PLOT
haz<-ggplot(hazardData,aes(factor(ID),HR))+ labs(x=NULL, y=NULL)
## RIGHT PANEL WITH LOG SCALE
rightPanel<-haz +
apply(hl_rows,1,function(x)annotation_custom(hl_rect(x["col"],alpha=0.4),as.numeric(x["ID"])-0.5,as.numeric(x["ID"])+0.5,-20,20)) +
geom_segment(aes(x = 2, y = 1, xend = 1.5, yend = 1)) +
geom_hline(aes(yintercept=1),linetype=2, linewidth=0.5)+
geom_boxplot(fill=boxColor,size=0.5, alpha=0.8)+
scale_y_log10() + coord_flip() +
geom_text(data=scaledata,aes(3,HR,label=HR), vjust=0.5, size=dataSize) +
geom_text(data=RtLabels,aes(x,y,label=lab, fontface="bold"), vjust=0.5, size=titleSize) +
geom_text(data=hrlabels,aes(factor(ID),4,label=lab),vjust=0.5, hjust=1, size=dataSize) +
geom_text(data=group_p,aes(factor(y),11,label=P, fontface="bold"),vjust=0.5, hjust=1, size=dataSize) +
geom_text(data=groupData,aes(factor(ID),6.5,label=P_S),vjust=0.5, hjust=1, size=dataSize) +
geom_text(data=LegendLabels,aes(x,y,label=lab, fontface="bold"),hjust=0.5, vjust=1, size=titleSize) +
geom_point(data=scaledata,aes(2.5,HR),shape=3,size=3) +
geom_point(aes(2,12),shape=3,alpha=0,vjust=0) +
geom_segment(aes(x = 2.5, y = 0, xend = 2.5, yend = 13)) +
geom_segment(aes(x = 2, y = 1, xend = 2, yend = 1.8),arrow=arrow(),linetype=1,size=1) +
geom_segment(aes(x = 2, y = 1, xend = 2, yend = 0.2),arrow=arrow(),linetype=1,size=1) +
theme_bare
## LEFT PANEL WITH NORMAL SCALE
leftPanel<-haz +
apply(hl_rows,1,function(x)annotation_custom(hl_rect(x["col"],alpha=0.4),as.numeric(x["ID"])-0.5,as.numeric(x["ID"])+0.5,-20,20)) +
coord_flip(ylim=c(0,5.5)) +
geom_point(aes(x=factor(ID),y=1),shape=3,alpha=0,vjust=0) +
geom_text(data=group_p,aes(factor(y),0.5,label=Group, fontface="bold"),vjust=0.5, hjust=0, size=dataSize) +
geom_text(data=groupData,aes(factor(ID),1,label=Subgroup),vjust=0.5, hjust=0, size=dataSize) +
geom_text(data=groupData,aes(factor(ID),5,label=NoP),vjust=0.5, hjust=1, size=dataSize) +
geom_text(data=LfLabels,aes(x,y,label=lab, fontface="bold"), vjust=0.5, hjust=0, size=4, size=titleSize) +
geom_segment(aes(x = 2.5, y = 0, xend = 2.5, yend = 5.5)) +
theme_bare
## PLOT THEM BOTH IN A GRID SO THEY MATCH UP
grid.arrange(leftPanel,rightPanel, widths=c(1,3), ncol=2, nrow=1)

The issue is that in your right panel plot you first map a continuous value on x via geom_segment and afterwards a discrete value via geom_boxplot. To fix that you could add a scale_x_discrete at the start of your code. Additionally I fixed the minor issues related to vjust in geom_point and a duplicated size argument in one of your geom_text layers.
## BASIC PLOT
haz <- ggplot(hazardData, aes(factor(ID), HR)) +
labs(x = NULL, y = NULL)
## RIGHT PANEL WITH LOG SCALE
rightPanel <- haz +
### Init the discrete x scale
scale_x_discrete() +
###
apply(hl_rows, 1, function(x) annotation_custom(hl_rect(x["col"], alpha = 0.4), as.numeric(x["ID"]) - 0.5, as.numeric(x["ID"]) + 0.5, -20, 20)) +
geom_segment(aes(x = 2, y = 1, xend = 1.5, yend = 1)) +
geom_hline(aes(yintercept = 1), linetype = 2, linewidth = 0.5) +
geom_boxplot(fill = boxColor, size = 0.5, alpha = 0.8) +
scale_y_log10() +
coord_flip() +
geom_text(data = scaledata, aes(3, HR, label = HR), vjust = 0.5, size = dataSize) +
geom_text(data = RtLabels, aes(x, y, label = lab, fontface = "bold"), vjust = 0.5, size = titleSize) +
geom_text(data = hrlabels, aes(factor(ID), 4, label = lab), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = group_p, aes(factor(y), 11, label = P, fontface = "bold"), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = groupData, aes(factor(ID), 6.5, label = P_S), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = LegendLabels, aes(x, y, label = lab, fontface = "bold"), hjust = 0.5, vjust = 1, size = titleSize) +
geom_point(data = scaledata, aes(2.5, HR), shape = 3, size = 3) +
geom_point(aes(2, 12), shape = 3, alpha = 0) +
geom_segment(aes(x = 2.5, y = 0, xend = 2.5, yend = 13)) +
geom_segment(aes(x = 2, y = 1, xend = 2, yend = 1.8), arrow = arrow(), linetype = 1, size = 1) +
geom_segment(aes(x = 2, y = 1, xend = 2, yend = 0.2), arrow = arrow(), linetype = 1, size = 1) +
theme_bare
## LEFT PANEL WITH NORMAL SCALE
leftPanel <- haz +
apply(hl_rows, 1, function(x) annotation_custom(hl_rect(x["col"], alpha = 0.4), as.numeric(x["ID"]) - 0.5, as.numeric(x["ID"]) + 0.5, -20, 20)) +
coord_flip(ylim = c(0, 5.5)) +
geom_point(aes(x = factor(ID), y = 1), shape = 3, alpha = 0) +
geom_text(data = group_p, aes(factor(y), 0.5, label = Group, fontface = "bold"), vjust = 0.5, hjust = 0, size = dataSize) +
geom_text(data = groupData, aes(factor(ID), 1, label = Subgroup), vjust = 0.5, hjust = 0, size = dataSize) +
geom_text(data = groupData, aes(factor(ID), 5, label = NoP), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = LfLabels, aes(x, y, label = lab, fontface = "bold"), vjust = 0.5, hjust = 0, size = titleSize) +
geom_segment(aes(x = 2.5, y = 0, xend = 2.5, yend = 5.5)) +
theme_bare
## PLOT THEM BOTH IN A GRID SO THEY MATCH UP
grid.arrange(leftPanel, rightPanel, widths = c(1, 3), ncol = 2, nrow = 1)
EDIT To get rid of the gap in your horizontal line and/or to extend the lines on the left and the right set yend=Inf and/or y=-Inf in the geom_segmentwhich draws the line.
library(gridExtra)
library(ggplot2)
## BASIC PLOT
haz <- ggplot(hazardData, aes(factor(ID), HR)) +
labs(x = NULL, y = NULL)
## RIGHT PANEL WITH LOG SCALE
rightPanel <- haz +
### Init the discrete x scale
scale_x_discrete() +
###
apply(hl_rows, 1, function(x) annotation_custom(hl_rect(x["col"], alpha = 0.4), as.numeric(x["ID"]) - 0.5, as.numeric(x["ID"]) + 0.5, -20, 20)) +
geom_segment(aes(x = 2, y = 1, xend = 1.5, yend = 1)) +
geom_hline(aes(yintercept = 1), linetype = 2, linewidth = 0.5) +
geom_boxplot(fill = boxColor, size = 0.5, alpha = 0.8) +
scale_y_log10() +
coord_flip() +
geom_text(data = scaledata, aes(3, HR, label = HR), vjust = 0.5, size = dataSize) +
geom_text(data = RtLabels, aes(x, y, label = lab, fontface = "bold"), vjust = 0.5, size = titleSize) +
geom_text(data = hrlabels, aes(factor(ID), 4, label = lab), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = group_p, aes(factor(y), 11, label = P, fontface = "bold"), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = groupData, aes(factor(ID), 6.5, label = P_S), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = LegendLabels, aes(x, y, label = lab, fontface = "bold"), hjust = 0.5, vjust = 1, size = titleSize) +
geom_point(data = scaledata, aes(2.5, HR), shape = 3, size = 3) +
geom_point(aes(2, 12), shape = 3, alpha = 0) +
geom_segment(aes(x = 2.5, y = 0, xend = 2.5, yend = Inf)) +
geom_segment(aes(x = 2, y = 1, xend = 2, yend = 1.8), arrow = arrow(), linetype = 1, size = 1) +
geom_segment(aes(x = 2, y = 1, xend = 2, yend = 0.2), arrow = arrow(), linetype = 1, size = 1) +
theme_bare
## LEFT PANEL WITH NORMAL SCALE
leftPanel <- haz +
apply(hl_rows, 1, function(x) annotation_custom(hl_rect(x["col"], alpha = 0.4), as.numeric(x["ID"]) - 0.5, as.numeric(x["ID"]) + 0.5, -20, 20)) +
coord_flip(ylim = c(0, 5.5)) +
geom_point(aes(x = factor(ID), y = 1), shape = 3, alpha = 0) +
geom_text(data = group_p, aes(factor(y), 0.5, label = Group, fontface = "bold"), vjust = 0.5, hjust = 0, size = dataSize) +
geom_text(data = groupData, aes(factor(ID), 1, label = Subgroup), vjust = 0.5, hjust = 0, size = dataSize) +
geom_text(data = groupData, aes(factor(ID), 5, label = NoP), vjust = 0.5, hjust = 1, size = dataSize) +
geom_text(data = LfLabels, aes(x, y, label = lab, fontface = "bold"), vjust = 0.5, hjust = 0, size = titleSize) +
geom_segment(aes(x = 2.5, y = -Inf, xend = 2.5, yend = Inf)) +
theme_bare
## PLOT THEM BOTH IN A GRID SO THEY MATCH UP
grid.arrange(leftPanel, rightPanel, widths = c(1, 3), ncol = 2, nrow = 1)

Related

Hide specific facet axis labels in ggplot

you'll see with the code below that I end up with a nicely faceted plot that looks how I need it, but all I want is to hide the y axis labels for all facets except the ones on the far left. So hide labels for facet 2, 3, 4, 6, and 7. That way I am just left with "White", "Black", and "Hispanic" on the far left of each row (I can clean up the prefix_ later). Any ideas?
d2 %>%
ggplot(., aes(x = var_new, y = coef,
ymin = ci_lower, ymax = ci_upper)) +
geom_point(color = "red") +
geom_errorbar(width = 0,
size = 1,
color = "red") +
facet_wrap(~model,
nrow = 2,
scales = "free") +
geom_hline(yintercept = 0, linetype = "dashed", color = "black", size = .3) +
coord_flip() +
theme_minimal(base_size = 10) +
theme(legend.position = "none")
structure(list(model = c(7, 6, 5, 7, 6, 5, 7, 6, 5, 4, 3, 4,
3, 4, 3, 2, 1, 2, 1, 2, 1), race = c("hispanic", "hispanic",
"hispanic", "black", "black", "black", "white", "white", "white",
"hispanic", "hispanic", "black", "black", "white", "white", "hispanic",
"hispanic", "black", "black", "white", "white"), var_new = c("ela_hispanic",
"math_hispanic", "sci_hispanic", "ela_black", "math_black", "sci_black",
"ela_white", "math_white", "sci_white", "after_hispanic", "before_hispanic",
"after_black", "before_black", "after_white", "before_white",
"part_hispanic", "full_hispanic", "part_black", "full_black",
"part_white", "full_white"), coef = c(0.91, 0.2615005, -0.0622102,
3.1966945, 0.9665615, 0.4419779, -4.1608082, -1.75, -3.4185874,
-1.72661788, -1.87514649, 0.61605887, 0.58634364, 0.87, 0.4,
1.52820746, 1.35976557, 1.08885352, 0.8323809019, 0.728991331,
1.53140561), ci_lower = c(0.3, -1.04316665, -1.68479242, -1.0382233,
-0.70264707, -1.29579134, -12.008101, -3, -6.4522842, -1.9858909,
-2.10047863, 0.41173674, 0.37007869, -0.3428254, -0.1, 1.21339829,
1.07813362, 0.778488586, 0.44183285, 0.30081336, 0.98770764),
ci_upper = c(1.2, 1.748, 1.560372, 7.4316126, 2.63577, 2.179747,
3.6864845, 0.01, -0.3848905, -1.467344828, -1.64981433, 0.8203809961,
0.802608596, 0.4, 0.8, 1.8430166, 1.64139752, 1.39921842,
1.22292898, 1.15716932, 2.0751036)), row.names = c(NA, -21L
), class = c("tbl_df", "tbl", "data.frame"))
I don't understand why folks continue to switch the x and y axis variables then use coord_flip to put them round the right way. This is confusing, unnecessary, and requires more code. It's best to just put the variables round the right way and keep the coord as-is.
Once that's done, the simplest solution is to put race on the y axis, and change scales to free_x. I've added a border around each panel to make things a bit clearer.
library(tidyverse)
ggplot(d2, aes(y = race, x = coef, xmin = ci_lower, xmax = ci_upper)) +
geom_errorbar(width = 0, linewidth = 1.5, color = "red3", alpha = 0.5) +
geom_point(shape = 21, fill = "red2", size = 3, color = 'white') +
facet_wrap(~ model, nrow = 2, scales = 'free_x') +
geom_vline(xintercept = 0, linetype = "dashed", linewidth = 0.3) +
theme_minimal(base_size = 14) +
theme(legend.position = "none",
panel.grid.major.y = element_blank(),
panel.border = element_rect(color = 'gray75', fill = NA))
If you want to include the prefixes in the facet titles (since they have a 1:1 correspondence with model), you could use tidyr::separate:
d2 %>%
separate(var_new, into = c('model_name', 'race')) %>%
mutate(model = paste(model, model_name, sep = ' - ')) %>%
ggplot(aes(y = race, x = coef, xmin = ci_lower, xmax = ci_upper)) +
geom_errorbar(width = 0, linewidth = 1.5, color = "red3", alpha = 0.5) +
geom_point(shape = 21, fill = "red2", size = 3, color = 'white') +
facet_wrap(~model, nrow = 2, scales = 'free_x') +
geom_vline(xintercept = 0, linetype = "dashed", linewidth = 0.3) +
theme_minimal(base_size = 14) +
theme(legend.position = "none",
panel.grid.major.y = element_blank(),
panel.border = element_rect(color = 'gray75', fill = NA))
Addendum
To compare coefficients across groups like this, it is normally better to put them all in a single linerange plot (similar to a forest plot). I think this provides a much better visualization that requires less cognitive effort from the reader. This also shows a good use-case for coord_flip, namely when you want a vertical dodge between groups.
d2 %>%
separate(var_new, into = c('model_name', 'race')) %>%
mutate(model = paste0('Model ', model, ' : ', model_name)) %>%
ggplot(aes(x = model, y = coef, ymin = ci_lower, ymax = ci_upper,
color = race)) +
annotate("segment", y = rep(-Inf, 3), yend = rep(Inf, 3),
x = c('Model 2 : part', 'Model 4 : after', 'Model 6 : math'),
xend = c('Model 2 : part', 'Model 4 : after', 'Model 6 : math'),
linewidth = 22, alpha = 0.05) +
coord_flip() +
geom_errorbar(width = 0, linewidth = 1, alpha = 0.5,
position = position_dodge(width = 0.5)) +
geom_point(size = 1.5, position = position_dodge(width = 0.5)) +
geom_hline(yintercept = 0, linetype = "dashed", color = "black",
linewidth = 0.3) +
scale_color_brewer(palette = 'Set1') +
theme_minimal(base_size = 14) +
guides(color = guide_legend(reverse = TRUE)) +
theme(panel.grid.major.y = element_blank(),
panel.border = element_rect(color = 'gray75', fill = NA),
axis.text.y = element_text(hjust = 0))

Recreate a plot without data

Is there a way to create a figure similar to the one below without having any data on this?
You could do something like this. You could add another geom_curve and a couple of geom_vlines.
library(tidyverse)
ggplot() +
geom_abline() +
geom_curve(aes(x = 0, y = 0, xend = 1, yend = 1), curvature = -0.4) +
annotate("text", x = 0.5, y = 0.5, label = "Line of Equality", angle = 35, vjust = 2) +
labs(x = "Individuals Neighbourhoods\nAcross Space", y = "Scoioeconomic Position") +
theme_minimal() +
theme(axis.text = element_blank())
Created on 2022-04-27 by the reprex package (v2.0.1)
For the arrows you can use this code: arrow = arrow(length = unit(0.5, "cm")) in a geom_segment. It is a bit tricky without any numbers, but maybe you want something like this:
library(ggplot2)
ggplot() +
geom_abline(slope = 1) +
geom_curve(aes(x = 0, y = 0, xend = 1, yend = 1), curvature = -0.4) +
geom_curve(aes(x = 0, y = 1, xend = 1, yend = 0.1), curvature = 0.4, linetype = "dashed") +
geom_segment(aes(x=0.9,y=0.98,xend=0.9,yend=0.12), arrow = arrow(length = unit(0.5, "cm"))) +
geom_segment(aes(x=0.02,y=0.85,xend=0.02,yend=0.23), arrow = arrow(length = unit(0.5, "cm"))) +
annotate("text", x = 0.5, y = 0.5, label = "Line of Equality", angle = 45, vjust = 2) +
annotate("text", x = 0.25, y = 0.75, label = "Income (etc.)", angle = 45, vjust = 2) +
annotate("text", x = 1.2, y = 0.1, label = "Corresponding\nExposure", angle = 0) +
labs(x = "Individuals Neighbourhoods\nAcross Space", y = "Scoioeconomic Position") +
scale_x_continuous(limits = c(0, 1.3)) +
theme_minimal() +
theme(axis.text = element_blank())
Output:

Add id labels to points above the limit line in ggplot

I have a dataframe like this
id <- c(5738180,51845,167774,517814,1344920)
amount <- c(3.76765976,0.85195407,1.96821355,0.01464609,0.57378284)
outlier <- c("TRUE","FALSE","FALSE","FALSE","FALSE")
df.sample <- data.frame(id,amount,outlier)
I am trying to plot the points and add an id label to any point that is above the limit. In this case (id=5738180)
I am plotting like this
library(tidyverse)
library(ggplot2)
library(ggrepel) # help avoid overlapping text labels
library(gridExtra) # adds custom table inside ggplot
library(scales) # breaks and labels for axes and legends
library(ggthemes) # Adding desired ggplot themes
df.sample %>%
ggplot(aes(x = as.numeric(row.names(df.sample)),
y = amount, label = as.character(id))) +
geom_point(alpha = 0.6, position = position_jitter(w = 0.05, h = 0.0),
aes(colour = (amount < 3)), size = 2) +
geom_hline(aes(yintercept = 3, linetype = "Limit"),
color = 'black', size = 1) +
geom_text(aes(y = 3,x = amount[4],
label = paste("Limit = ", round(3, 3)),
hjust = 0, vjust = 1.5)) +
geom_text_repel(data = subset(df.sample, outlier == 'TRUE'),
nudge_y = 0.75,
size = 4,
box.padding = 1.5,
point.padding = 0.5,
force = 100,
segment.size = 0.2,
segment.color = "grey50",
direction = "x") +
geom_label_repel(data = subset(df.sample, outlier == 'TRUE'),
nudge_y = 0.75,
size = 4,
box.padding = 0.5,
point.padding = 0.5,
force = 100,
segment.size = 0.2,
segment.color = "grey50",
direction = "x")
labs(title = "Outlier Detection",
y = "amount",
x = "") +
theme_few() +
theme(legend.position = "none",
axis.text = element_text(size = 10, face = "bold"),
axis.title = element_text(size = 10, face = "bold"),
plot.title = element_text(colour = "blue", hjust = 0.5,
size = 15, face = "bold"),
strip.text = element_text(size = 10, face = "bold")) +
scale_colour_manual(values = c("TRUE" = "green","FALSE" = "red"))
I am running into an error "Error: Aesthetics must be either length 1 or the same as the data (1): x"
Can someone point me in the right direction?
The issue is with geom_text_repel() and geom_label_repel(). You subset the data, which now only includes 1 row, but the aes() are inheriting from the original data which have 5 rows, hence the error. To fix this, subset the data outside of the ggplot() call, and change the aesthetics for it. You are also missing a + after geom_label_repel() and the result below modifies the nudge_y to nudge_x and removes the geom_text_repel().
outliers <- subset(df.sample, outlier == TRUE)
ggplot(data = df.sample,
aes(x = as.numeric(row.names(df.sample)),
y = amount,
label = as.character(id))) +
geom_point(alpha = 0.6,
position = position_jitter(w = 0.05, h = 0.0),
aes(colour = (amount < 3)),
size = 2) +
geom_hline(aes(yintercept = 3,
linetype = "Limit"),
color = 'black',
size = 1) +
geom_text(aes(y = 3,x = amount[4],
label = paste("Limit = ",
round(3, 3)),
hjust = 0,
vjust = 1.5)) +
geom_label_repel(data = outliers,
aes(x = as.numeric(rownames(outliers)),
y = amount,
label = amount),
nudge_x = 0.75,
size = 4,
box.padding = 0.5,
point.padding = 0.5,
force = 100,
segment.size = 0.2,
segment.color = "grey50",
direction = "x",
inherit.aes = F) +
labs(title = "Outlier Detection",
y = "amount",
x = "") +
theme_few() +
theme(legend.position = "none",
axis.text = element_text(size = 10, face = "bold"),
axis.title = element_text(size = 10, face = "bold"),
plot.title = element_text(colour = "blue", hjust = 0.5,
size = 15, face = "bold"),
strip.text = element_text(size = 10, face = "bold")) +
scale_colour_manual(values = c("TRUE" = "green","FALSE" = "red"))

annotate font is different from the other texts in the figure with ggplot2

I am trying to add some annotations to the graph that I create with ggplot2 2. The font size and type seem to be different for the annotations compared to everything else. I have a font size of 20 for the labels and the legend, but even with a size of 12 and specifying the font type (family) as Arial, the annotations look very different than the rest of the graph.
fontsize <- 20 / .pt
plot <- ggplot(material) + theme_void() + scale_color_hue() +
aes(x = strain, y = stress, colour = type, group = type) +
geom_xspline(size = 1, spline_shape = 0.5) + labs(x = "Strain (in/in)",
y = "Stress (ksi)", color = element_blank()) + theme(legend.position = "bottom",
legend.box.background = element_rect(colour = "black", size = 0.5),
axis.title.y = element_text(size=20, colour="black", angle = 90, vjust = -22, hjust = 0.55),
axis.title.x = element_text(size=20, colour="black", vjust = 7, hjust = 0.6),
axis.text.y = element_blank(), axis.text.x = element_blank(),
legend.text=element_text(size=20), legend.box.margin = margin(10,10,10,10),
legend.spacing.x = unit(0, 'cm')) +
scale_y_continuous(trans = "reverse") + scale_x_continuous(trans = "reverse") +
geom_vline(aes(xintercept = 0), size = 0.3) + geom_hline(aes(yintercept = 0), size = 0.3) +
geom_segment(aes(x = 0, y = 0, xend = -0.0009226, yend = 0), size = 1) +
geom_segment(aes(x = -0.0009226, y = 0, xend = -0.0011707, yend = -0.94), size = 1) +
geom_segment(aes(x = 0, y = -4.958, xend = -0.00333, yend = -4.958),
size = 0.6, linetype = "dotted", colour = "black") +
geom_segment(aes(x = -0.00333, y = 0, xend = -0.00333, yend = -4.958),
size = 0.6, linetype = "dotted", colour = "black") +
geom_segment(aes(x = -0.0024, y = 0, xend = -0.0024, yend = -4.958),
size = 0.6, linetype = "dotted", colour = "black") +
geom_segment(aes(x = -3.75E-03, y = 0, xend = -3.75E-03, yend = -4.6011),
size = 0.6, linetype = "dotted", colour = "black") +
geom_segment(aes(x = -0.0046736, y = 0, xend = -0.0046736, yend = -4.6011),
size = 0.6, linetype = "dotted", colour = "black") +
annotate("text", x = 0, y= -0.94, label = "f'", size = fontsize, family = "sans")
The resulted graph is
How can I send all the text in the figure to the same font and size?
EDIT
The issue is that the font size in the theme is usually specified in pt (points) while when specifying size inside of the geoms or annotate it is measured in mm. This means that the size of you axis.title is 20pt, while the size of your annotation is 12mm or approx. 34pt. Therefore, if you want the same font sizes you have to do a conversion for which ggplot2 offers a helper constant .pt.
If you want the annotation to have the same size as e.g. the axis.title you have to set the fontsize in annotate equal to 20 / .pt.
library(ggplot2)
family <- "sans"
fontsize <- 20 / .pt
ggplot(mtcars, aes(hp, mpg)) + theme_void() +
geom_point(aes(color = factor(cyl))) +
annotate("text", x = 0, y= -0.94, label = "f'", size = fontsize, family = "sans") +
geom_text(data = data.frame(hp = 100, mpg = -.94), aes(label = "f'"), size = fontsize) +
theme(legend.position = "bottom",
legend.box.background = element_rect(colour = "black", size = 0.5),
axis.title.y = element_text(size=20, colour="black", angle = 90, vjust = -22, hjust = 0.55),
axis.title.x = element_text(size=20, colour="black", vjust = 7, hjust = 0.6),
axis.text.y = element_blank(), axis.text.x = element_blank(),
legend.text=element_text(size=20), legend.box.margin = margin(10,10,10,10),
legend.spacing.x = unit(0, 'cm'))

Exact Positioning of multiple plots in ggplot2 with grid.arrange

I'm trying to create a multiple plot with the same x-axis but different y-axes, because I have values for two groups with different ranges. As I want to control the values of the axes (respectively the y-axes shall reach from 2.000.000 to 4.000.000 and from 250.000 to 500.000), I don't get along with facet_grid with scales = "free".
So what I've tried is to create two plots (named "plots.treat" and "plot.control") and combine them with grid.arrange and arrangeGrob. My problem is, that I don't know how to control the exact position of the two plots, so that both y-axes are positioned on one vertical line. So in the example below the second plot's y-axis needs to be positioned a bit more to the right.
Here is the code:
# Load Packages
library(ggplot2)
library(grid)
library(gridExtra)
# Create Data
data.treat <- data.frame(seq(2005.5, 2015.5, 1), rep("SIFI", 11),
c(2230773, 2287162, 2326435, 2553602, 2829325, 3372657, 3512437,
3533884, 3519026, 3566553, 3527153))
colnames(data.treat) <- c("Jahr", "treatment",
"Aggregierte Depositen (in Tausend US$)")
data.control <- data.frame(seq(2005.5, 2015.5, 1), rep("Nicht-SIFI", 11),
c(324582, 345245, 364592, 360006, 363677, 384674, 369007,
343893, 333370, 318409, 313853))
colnames(data.control) <- c("Jahr", "treatment",
"Aggregierte Depositen (in Tausend US$)")
# Create Plot for data.treat
plot.treat <- ggplot() +
geom_line(data = data.treat,
aes(x = `Jahr`,
y = `Aggregierte Depositen (in Tausend US$)`),
size = 1,
linetype = "dashed") +
geom_point(data = data.treat,
aes(x = `Jahr`,
y = `Aggregierte Depositen (in Tausend US$)`),
fill = "white",
size = 2,
shape = 24) +
scale_x_continuous(breaks = seq(2005, 2015.5, 1),
minor_breaks = seq(2005, 2015.5, 0.5),
limits = c(2005, 2015.8),
expand = c(0.01, 0.01)) +
scale_y_continuous(breaks = seq(2000000, 4000000, 500000),
minor_breaks = seq(2000000, 4000000, 250000),
labels = c("2.000.000", "2.500.000", "3.000.000",
"3.500.000", "4.000.000"),
limits = c(2000000, 4000000),
expand = c(0, 0.01)) +
theme(text = element_text(family = "Times"),
axis.title.x = element_blank(),
axis.title.y = element_blank(),
axis.line.x = element_line(color="black", size = 0.6),
axis.line.y = element_line(color="black", size = 0.6),
legend.position = "none") +
geom_segment(aes(x = c(2008.7068),
y = c(2000000),
xend = c(2008.7068),
yend = c(3750000)),
linetype = "dotted") +
annotate(geom = "text", x = 2008.7068, y = 3875000, label = "Lehman\nBrothers + TARP",
colour = "black", size = 3, family = "Times") +
geom_segment(aes(x = c(2010.5507),
y = c(2000000),
xend = c(2010.5507),
yend = c(3750000)),
linetype = "dotted") +
annotate(geom = "text", x = 2010.5507, y = 3875000, label = "Dodd-Frank-\nAct",
colour = "black", size = 3, family = "Times") +
geom_rect(aes(xmin = 2007.6027, xmax = 2009.5, ymin = -Inf, ymax = Inf),
fill="dark grey", alpha = 0.2)
# Create Plot for data.control
plot.control <- ggplot() +
geom_line(data = data.control,
aes(x = `Jahr`,
y = `Aggregierte Depositen (in Tausend US$)`),
size = 1,
linetype = "solid") +
geom_point(data = data.control,
aes(x = `Jahr`,
y = `Aggregierte Depositen (in Tausend US$)`),
fill = "white",
size = 2,
shape = 21) +
scale_x_continuous(breaks = seq(2005, 2015.5, 1), # x-Achse
minor_breaks = seq(2005, 2015.5, 0.5),
limits = c(2005, 2015.8),
expand = c(0.01, 0.01)) +
scale_y_continuous(breaks = seq(250000, 500000, 50000),
minor_breaks = seq(250000, 500000, 25000),
labels = c("250.000", "300.000", "350.000", "400.000",
"450.000", "500.000"),
limits = c(250000, 500000),
expand = c(0, 0.01)) +
theme(text = element_text(family = "Times"),
axis.title.x = element_blank(), # Achse
axis.title.y = element_blank(), # Achse
axis.line.x = element_line(color="black", size = 0.6),
axis.line.y = element_line(color="black", size = 0.6),
legend.position = "none") +
geom_segment(aes(x = c(2008.7068),
y = c(250000),
xend = c(2008.7068),
yend = c(468750)),
linetype = "dotted") +
annotate(geom = "text", x = 2008.7068, y = 484375, label = "Lehman\nBrothers + TARP",
colour = "black", size = 3, family = "Times") +
geom_segment(aes(x = c(2010.5507),
y = c(250000),
xend = c(2010.5507),
yend = c(468750)),
linetype = "dotted") +
annotate(geom = "text", x = 2010.5507, y = 484375, label = "Dodd-Frank-\nAct",
colour = "black", size = 3, family = "Times") +
geom_rect(aes(xmin = 2007.6027, xmax = 2009.5, ymin = -Inf, ymax = Inf),
fill="dark grey", alpha = 0.2)
# Combine both Plots with grid.arrange
grid.arrange(arrangeGrob(plot.treat, plot.control,
ncol = 1,
left = textGrob("Aggregierte Depositen (in Tausend US$)",
rot = 90,
vjust = 1,
gp = gpar(fontfamily = "Times",
size = 12,
colout = "black",
fontface = "bold")),
bottom = textGrob("Jahr",
vjust = 0.1,
hjust = 0.2,
gp = gpar(fontfamily = "Times",
size = 12,
colout = "black",
fontface = "bold"))))
Do:
install.packages("cowplot")
but do not library(cowplot) as it'll mess up your theme work.
Then, do:
grid.arrange(
arrangeGrob(cowplot::plot_grid(plot.treat, plot.control, align = "v", ncol=1),
ncol = 1,
left = textGrob("Aggregierte Depositen (in Tausend US$)",
rot = 90,
vjust = 1,
gp = gpar(fontfamily = "Times",
size = 12,
colout = "black",
fontface = "bold")),
bottom = textGrob("Jahr",
vjust = 0.1,
hjust = 0.2,
gp = gpar(fontfamily = "Times",
size = 12,
colout = "black",
fontface = "bold"))))

Resources