I have the following reproducible code which gets me the plot listed below:
require(ggplot2)
set.seed(123)
ChickWt <- data.frame(ChickWeight, AR = sample(c("p=0", "p=1", "hat(p)"), size = 578, replace = T))
exprvec <- expression( p==hat(p), p==0, p==1)
p1 <- ggplot(ChickWt, aes(x=Time, y=weight, colour=Diet, Group = Chick, linetype = AR)) + geom_line()
p1 <- p1 + scale_linetype_manual(values=c(2,4,1), labels = exprvec,name="AR order") + theme_bw() + theme(legend.justification=c(1,-0.2), legend.position=c(0.3,0.2), legend.text=element_text(size=10), legend.title=element_text(size=10), axis.title.x=element_text(size=10), axis.title.y=element_text(size = 10), legend.key = element_blank(), legend.background = element_rect(color="black",size = 0.1)) + ylim(c(0,400)) + guides(fill=guide_legend(ncol=2))
but I would like the legend on Diet and AR order in two separate columns. How do I get this to work? Clearly, the guides(fill=guide_legend(ncol=2)) has no effect, perhaps because these are two separate legends.
Thanks for suggestions!
The reason that guides(fill=guide_legend(ncol=2)) does not work is because it only refers to the fill-legend and not to the linetype-legend. You can position the legends next to each other by using legend.box = "horizontal":
ggplot(ChickWt, aes(x=Time, y=weight, colour=Diet, Group = Chick, linetype = AR)) +
geom_line() +
scale_linetype_manual(values=c(2,4,1), labels = exprvec,name="AR order") +
theme_bw() +
theme(legend.justification=c(1,-0.2),
legend.position=c(0.3,0.2),
legend.text=element_text(size=10),
legend.title=element_text(size=10),
axis.title.x=element_text(size=10),
axis.title.y=element_text(size = 10),
legend.key = element_blank(),
legend.background = element_rect(color="black",size = 0.1),
legend.box = "horizontal") +
ylim(c(0,400))
which gives:
Related
I'm trying to recreate a bar graph found on page 4 of the following report:
The figure has three bars with the first two stacked and the third dodged next to it. I've seen iterations of this question but none that recreate the figure in this exact way.
Here is the data:
a <- rep(c('RHB', 'FERS', 'CSRS'), 3)
b <- c(rep('Assets', 3), rep('Amount Past Due', 3),
rep('Actuarial Liability', 3))
c <- c(45.0, 122.5, 152.3, 47.2, 3.4, 4.8, 114.4, 143.4, 181.3)
df <- data.frame(a,b,c)
names(df) <- c('Fund', 'Condition', 'Value')
And what I've managed so far:
p <- ggplot(subset_data, aes(fill=Condition, y=Value, x=Fund)) +
geom_bar(position="stack", stat="identity") +
coord_flip()
I'm not partial to ggplot so if there's another tool that works better I'm ok using another package.
Taking some ideas from the link #aosmith posted.
You can call geom_bar twice, once with Assets and Amounts Past Due stacked, and again with just Actuarial Liability.
You can use width to make the bars thinner, then nudge one set of bars so the two geom_bar calls are not overlapping. I chose to make the width 0.3 and nudge by 0.3 so the edges just line up. If you nudge by more you will see a gap between the two bars.
Edit: add some more formatting and numeric labels
library(tidyverse)
library(scales)
df_al <- filter(df, Condition == 'Actuarial Liability')
df_xal <- filter(df, Condition != 'Actuarial Liability')
bar_width <- 0.3
hjust_lab <- 1.1
hjust_lab_small <- -0.2 # hjust for labels on small bars
ggplot() +
theme_classic() +
geom_bar(data = df_al,
aes(fill=Condition, y=Value, x=Fund),
position = position_nudge(x = -bar_width),
width = bar_width,
stat="identity") +
geom_bar(data = df_xal,
aes(fill=Condition, y=Value, x=Fund),
position="stack",
stat="identity",
width = bar_width) +
geom_text(data = df_al,
aes(label= dollar(Value, drop0trailing = TRUE), y=Value, x=Fund),
position = position_nudge(x = -bar_width),
hjust = hjust_lab) +
geom_text(data = df_xal,
aes(label= dollar(Value, drop0trailing = TRUE), y=Value, x=Fund),
position="stack",
hjust = ifelse(df_xal$Value < 5, hjust_lab_small, hjust_lab)) +
scale_fill_manual(values = c('firebrick3', 'lightsalmon', 'dodgerblue')) +
scale_y_continuous(breaks = seq(0,180, by = 20), labels = dollar) +
coord_flip() +
labs(x = NULL, y = NULL, fill = NULL) +
theme(legend.position = "bottom")
I think I would use the "sneaky facet" method, after adding a dummy variable to dodge the columns and making Fund a factor with the correct order:
df$not_liability <-df$Condition != "Actuarial Liability"
df$Fund <- factor(df$Fund, levels = c('RHB', 'FERS', 'CSRS'))
Most of the plotting code is then an attempt to copy the look of the supplied plot:
ggplot(df, aes(fill=Condition, y=Value, x=not_liability)) +
geom_bar(position = "stack", stat = "identity") +
scale_x_discrete(expand = c(0.5, 0.5)) +
scale_y_continuous(breaks = 0:10 * 20, labels = scales::dollar) +
coord_flip() +
facet_grid(Fund~., switch = "y") +
scale_fill_manual(values = c("#c00000", "#f7c290", "#0071bf"), name = "") +
theme_classic() +
theme(panel.spacing = unit(0, "points"),
strip.background = element_blank(),
axis.text.y = element_blank(),
axis.ticks.length.y = unit(0, "points"),
axis.title = element_blank(),
strip.placement = "outside",
strip.text = element_text(),
legend.position = "bottom",
panel.grid.major.x = element_line())
I'd like to add the count values calculated in the geom_histogram function on ggplot2. I've put the ggplot2 into a loop so I can produce multiple plots, in my case 30 but for ease, here is a dummy set for only four plots. Facet wrap didn't work as the geom density was pooling the data across all factors before calculating proportions, rather than within a factor/variable. To produce this plot, I've essentially mixed a whole bunch of code from various sources so credit to them.
library(dplyr)
library(ggplot2)
library(ggridges)
library(reshape2)
library(gridExtra)
#Make the data#
df.fact <- data.frame("A"=rnorm(400, mean = 350, sd=160),"B"=rnorm(400, mean = 300, sd=100), "C"=rnorm(400, mean = 200, sd=80), names=rep(factor(LETTERS[23:26]), 100))
df.test<-melt(df.fact, id.vars = "names", value.name = "Length2")
names(df.test)[names(df.test) =="variable"] <- "TSM.FACT"
#Create the plotlist##
myplots <- list()
#Loop for plots##
for(i in 1:(length(unique(df.test$names)))){
p1 <- eval(substitute(
ggplot(data=df.test[df.test$names == levels(df.test$names)[i],], aes(x=Length2, group=TSM.FACT, colour = TSM.FACT, fill=TSM.FACT)) +
geom_histogram(aes( y = stat(width*density)), position = "dodge", binwidth = 50, alpha =0.4, show.legend=T)+
ggtitle(paste0(levels(df.test$names)[i]))+
geom_density_line(stat="density", aes(y=(..count..)/sum(..count..)*50), alpha=0.3, size=0.5, show.legend=F) +
geom_vline(data=ddply(df.test[df.test$names == levels(df.test$names)[i],], ~ TSM.FACT, numcolwise(mean)), mapping=aes(xintercept = Length2, group=TSM.FACT, colour=TSM.FACT), linetype=2, size=1, show.legend=F) +
scale_y_continuous(labels = percent_format()) +
ylab("relative frequency") +
scale_color_manual(values= c("#00B2EE", "#1E90FF", "#104E8B")) +
scale_fill_manual(values= c("#00B2EE", "#1E90FF", "#104E8B")) +
theme_bw() + theme(
plot.title = element_text(lineheight=0.5, hjust= 0.5, size=10),
strip.text.y = element_text(hjust = 1, angle = 0),
strip.text.x = element_text(size=10, vjust = 0.9),
strip.text=element_text(margin = margin(t=0.3,r=1,b=0.3,l=1), size=8, debug = F, vjust=0.2),
strip.background = element_blank(),
axis.text.x = element_text(size=8, angle=0, vjust=0.2, margin = margin(t=0.3,r=0.1,b=0.3,l=0.1)),
axis.title.x=element_blank(),
axis.title.y=element_blank(),
axis.line.x=element_line(colour="black"),
axis.line.y=element_line(colour="black"),
panel.grid.minor = element_blank(),
panel.border=element_blank(),
panel.background=element_blank(),
legend.position=(c(0.9,0.9)),
legend.title = element_blank(),
legend.key = element_blank()),
list(i = i)))
print(i)
print(p1)
myplots[[i]] <- p1
plot(p1)
}
#Join the plots
panelplot=grid.arrange(plotlist = myplots, grobs = myplots, shared.legend=T)
Unfortunately I am unable to reproduce your example. I can recommend adding a column that includes the sum of each bar (let's name it "Bar")
The required addition to the ggplot code then involves:
geom_text(aes(label = Bar), position = position_stack(vjust = 1)) +
The text height above the bar can be adjusted with vjust
I am preparing a multiplot for with two y axes for Inc and Ratio.
I distinguished each plot with different colours to represent three regions.
There are three items I am not successful at:
I do have now, two lines with the same colour in each plot. I would like to change one of them to be dashed (Ratio one).
I need to add SE bars to Inc line (from Inc column)
I would like to add geom_points() so there are also points at the nodes where lines are connecting, only for aesthetic reasons.
This is as far as I get:
df <- data.frame(c(2009,2009,2009,2009,2010,2010,2010,2010,2011,2011,2011,2011,
2012,2012,2012,2012,2013,2013,2013,2013),
c("N","S","W","W","N","S","W","W","N","S","W","W","N","S","W","W",
"N","S","W","W"),
c("Luo","Aka","Opo","Mya","Luo","Aka","Opo","Mya",
"Luo","Aka","Opo","Mya","Luo","Aka","Opo","Mya",
"Luo","Aka","Opo","Mya"),
runif(20,0,1),runif(20,0,1),
runif(20,0,0.1))
colnames(df) <- c("Year","Region","District","Inc","Ratio","Inc_SE")
# Order of drawing in facet
df$District<- factor(df$District,
levels = c("Opo",
"Mya",
"Luo",
"Aka"))
p <- ggplot(data=df, aes(x = Year))
p <- p + geom_line(aes(y = Inc))
p <- ggplot(df, aes(x = Year, y=df$Inc))
p <- p + geom_line(aes(y = Inc))
p <- ggplot(df, aes(x = Year))
p <- p + geom_line(aes(y = Inc, colour = Region))
p <- p + theme_bw()+
theme(plot.title = element_text(hjust = 1))+
theme(legend.position="none")+
theme(axis.title.x = element_text(face ="bold", colour="black", size=11),
axis.text.x = element_text(angle=90, vjust=0.5, size=7, family = "serif"),
axis.title.y = element_text(face = "bold", colour = "black", size=10))
# adding Ratio
p <- p + geom_line(aes(y = Ratio, colour = Region,linetype = "dashed")) # here dashed is not recognised by R
# now adding the secondary axis
p <- p + scale_y_continuous(sec.axis = sec_axis(~.*1, name = "Ratio"))
p <- p + scale_colour_manual(values = c("blue", "red","black"))
p <- p +
theme_bw()+
theme(plot.title = element_text(hjust = 1))+
theme(legend.position="none")+
theme(axis.title.x = element_text(face ="bold", colour="black", size=11),
axis.text.x = element_text(angle=90, vjust=0.5, size=9, family = "serif"),
axis.title.y = element_text(face = "bold", colour = "black", size=10))
# Breaking down to separate graphs
p_facet = p + facet_wrap(~ df$District,
ncol = 2)
p_facet
You can try a tidyverse. The trick is to transform the data from wide to long (here I used gather). then you can easily add points, lines and Inc_SE as ribbon.
library(tidyverse)
df %>%
gather(k,v, -Year, -Region, -District, -Inc_SE) %>%
ggplot(aes(Year, v, group = k, color=Region, linetype=k)) +
geom_ribbon(data=. %>% filter( k == "Inc"),
aes(ymin=v-Inc_SE, ymax=v+Inc_SE),
alpha=0.2,color=NA,
show.legend = F) +
geom_line() +
geom_point(show.legend = F)+
scale_y_continuous(sec.axis = sec_axis(~.*1, name = "Ratio"))+
facet_wrap(~ District) +
labs(y="Inc") +
theme_bw() +
theme(legend.position = "bottom")
I am trying to create a barplot with two x-axis (grouped x-axis):
# read data
tmp <- read.table(text = "label CNV_x CNV_Type
17p -1 Loss
9p -1 Loss
16q 1 Gain
10p 1 Gain
8q 1 Gain
13q 1 Gain", header = T)
tmp$CNV_Type <- relevel(tmp$CNV_Type, ref = 'Loss')
# plot
ggplot(tmp, aes(x = label, y = CNV_x)) +
geom_bar(stat = 'identity') +
theme_bw() +
geom_hline(yintercept = 0) +
coord_flip() +
facet_wrap(~CNV_Type, strip.position = "bottom", scales = "free_x") +
theme(panel.spacing = unit(0, "lines"),
strip.background = element_blank(),
strip.placement = "outside",
panel.border = element_rect(colour = NA))
This creates a plot like this:
This plot shows 0.00 twice on x-axis and I can't figure out a way to remove the spacing between the two vertical lines separating the strips (one is Gain and other is Loss).
Any help would be much appreciated. Thanks!
UPDATE: I added scale_y_continuous(expand = c(0, 0)) as suggested below:
ggplot(tmp, aes(x = label, y = CNV_x)) +
geom_bar(stat = 'identity') +
theme_bw() +
geom_hline(yintercept = 0) +
scale_y_continuous(expand = c(0, 0)) +
coord_flip() +
facet_wrap(~CNV_Type, strip.position = "bottom", scales = "free_x") +
theme(panel.spacing = unit(0, "lines"),
strip.background = element_blank(),
strip.placement = "outside",
panel.border = element_rect(colour = NA))
This creates a plot like this:
The only issue now is there is no spacing between the bars and the left and right margins of the plot - not sure why that happened.
I would not use facets here. A couple of options. You could indicate the type by colour:
tmp %>%
ggplot(aes(label, CNV_x)) +
geom_col(aes(fill = CNV_Type)) +
geom_hline(yintercept = 0) +
coord_flip() +
scale_fill_manual(values = c("darkorange", "skyblue3"))
And/or add the labels for type to the plot using annotate. That requires some manual fiddling with x, y and expand to get it right:
tmp %>%
ggplot(aes(label, CNV_x)) +
geom_col() +
geom_hline(yintercept = 0) +
coord_flip() +
annotate("text",
label = c("Loss", "Gain"),
x = c(7, 7),
y = c(-0.5, 0.5)) +
scale_x_discrete(expand = c(0.1, 0.1))
I need to gather two facet columns into one column with ggplot2.
In the following example, I need to overlay the content of the two columns DEG and RAN into one, while giving different colours to DEG and RAN data (small points and smooth line) and provide the corresponding legend (so I can distinguish them as they are overlayed).
I feel my code is not too, too far from what I need, but the relative complexity of the dataset blocks me. How to go about achieving this in ggplot2?
Here's my code so far:
require(reshape2)
library(ggplot2)
library(RColorBrewer)
fileName = paste("./4.csv", sep = "") # csv file available here: https://www.dropbox.com/s/bm9hd0t5ak74k89/4.csv?dl=0
mydata = read.csv(fileName,sep=",", header=TRUE)
dataM = melt(mydata,c("id"))
dataM = cbind(dataM,colsplit(dataM$variable,pattern = "_",names = c("NM", "ORD", "CAT")))
dataM$variable <- NULL
dataM <- dcast(dataM, ... ~ CAT, value.var = "value")
my_palette <- colorRampPalette(rev(brewer.pal(11, "Spectral")))
ggplot(dataM, aes(x=NR ,y= ASPL)) +
geom_point(size = .4,alpha = .5) +
stat_smooth(se = FALSE, size = .5) +
theme_bw() +
theme(plot.background = element_blank(),
axis.line = element_blank(),
legend.key = element_blank(),
legend.title = element_blank()) +
scale_y_continuous("ASPL", expand=c(0,0), limits = c(1, 7)) +
scale_x_continuous("NR", expand=c(0,0), limits = c(0, 100)) +
theme(legend.position="bottom") +
theme(axis.title.x = element_text(vjust=-0.3, face="bold", size=12)) +
theme(axis.title.y = element_text(vjust=1.5, face="bold", size=12)) +
ggtitle("Title") + theme(plot.title = element_text(lineheight=.8, face="bold")) +
theme(title = element_text(vjust=2)) +
facet_grid(NM ~ ORD)
Here's what it gives me right now:
Extra question: how come DEG/SF doesn't show a smooth line?
You can use the group aesthetic to define that data points with the same value of ORD belong together. You can also map aesthetics shape and color to this variable. You can also use . to specify that the facets are not split along a specific dimension.
I have made the changes to your code below after transforming NR and ASPL to numeric variables:
dataM$NR <- as.integer(dataM$NR)
dataM$ASPL <- as.numeric(dataM$ASPL)
ggplot(dataM, aes(x=NR ,y= ASPL, group=ORD, color=ORD)) +
geom_point(size = .7,alpha = .5, aes(shape=ORD)) + ## increased size
stat_smooth(se = FALSE, size = .5) +
theme_bw() +
theme(plot.background = element_blank(),
axis.line = element_blank(),
legend.key = element_blank(),
legend.title = element_blank()) +
scale_y_continuous("ASPL", expand=c(0,0), limits = c(1, 7)) +
scale_x_continuous("NR", expand=c(0,0), limits = c(0, 100)) +
theme(legend.position="bottom") +
theme(axis.title.x = element_text(vjust=-0.3, face="bold", size=12)) +
theme(axis.title.y = element_text(vjust=1.5, face="bold", size=12)) +
ggtitle("Title") + theme(plot.title = element_text(lineheight=.8, face="bold")) +
theme(title = element_text(vjust=2)) +
facet_grid(NM ~.)