My plot currently looks like this:
I want to change the shape in the legend (which is currently "a") for only points that indicate the respective colours. This is my code so far:
ggplot(data=pca2.data, aes(x=X, y=Y, label=Sample, colour = col)) +
geom_text() +
xlab(paste("PC1 - ", pca2.var.per[1], "%", sep="")) +
ylab(paste("PC2 - ", pca2.var.per[2], "%", sep="")) +
theme_bw() +
ggtitle("My PCA Graph") +
geom_hline(yintercept = 0) +
geom_vline(xintercept = 0) +
scale_color_manual(values=c("black", "red", "green"), labels = c("No significant difference", "Sharpe Decrease", "Sharpe Increase")) +
theme(legend.position = 'bottom') + guides(color=guide_legend(""))
I already tried adding "shape = c(20, 20, 20)" inside of "guide_legend", but it changed nothing.
Just put an empty point layer and don't plot legend for geom_text
As you didn't provide data, I've used mtcars dataset but it should translate to your problem
ggplot(mtcars, aes(mpg, cyl, label=rownames(mtcars), color=factor(carb))) +geom_point(shape=NA)+
geom_text( show.legend = F ) + guides(colour=guide_legend(override.aes = list(shape = 16)))
Related
I have the following code which yields the figure below:
ggplot(data=data.frame(x=x, y=y, mass=mass)) +
geom_line(mapping = aes(x=x, y=y, linetype='Gompertz predicted mass', col='Gompertz predicted mass')) +
geom_point(mapping = aes(x=x, y=mass, shape='Actual mass',col='Actual mass')) +
theme_bw() +
ylab('Mass') +
xlab('t') +
scale_color_manual(name='',values = c("black",'red')) +
scale_linetype_manual(name='',values = c("solid")) +
scale_shape_manual(name='', values = c(19)) +
scale_x_continuous(breaks=seq(4,26,2)) +
ylim(c(0, 20000)) +
ggtitle('Problem 3: Plot of tumor mass with time')
Notice how the legend is separated. I'd like to merge it for shape and color. When the geoms are the same, the technique of using scale_something_manual works perfectly fine to merge the legends. However, I'm having trouble with it here since I have two different geoms.
The problem is similar to the one described in https://github.com/tidyverse/ggplot2/issues/3648. There is no elegant solution at the moment. Because you haven't included any data, I've presumed that your problem is conceptually similar to the plot below:
library(ggplot2)
ggplot(mpg, aes(displ, hwy)) +
geom_point(aes(shape = "Point", colour = "Point")) +
geom_smooth(aes(linetype = "Line", colour = "Line"),
formula = y ~ x, se = FALSE, method = "loess") +
scale_colour_manual(values = c("red", "black")) +
scale_linetype_manual(values = "solid") +
scale_shape_manual(values = 19)
The way to fix the problem is to get rid of the linetype and shape aesthetics and scales, and instead override aesthetics at the level of the legend.
ggplot(mpg, aes(displ, hwy)) +
geom_point(aes(colour = "Point")) +
geom_smooth(aes(colour = "Line"),
formula = y ~ x, se = FALSE, method = "loess") +
scale_colour_manual(
values = c("red", "black"),
guide = guide_legend(override.aes = list(shape = c(NA, 19),
linetype = c(1, NA)))
)
Created on 2021-09-04 by the reprex package (v2.0.1)
legend <- c("score" = "black", "answer" = "red")
plot <- df_l %>% ggplot(aes(date, score, color = "score")) + geom_line() +
geom_vline(aes(xintercept = getDate(df_all %>% filter(name == List[5])), color = "answer"), linetype = "dashed", size = 1,) +
scale_color_manual(name = "Legend", values = legend) +
scale_x_date(labels = date_format("%m/%y"), breaks = date_breaks("months")) +
theme(axis.text.x = element_text(angle=45)) +
labs(title = "", x = "", y = "", colors = "Legend")
I get the result above and could not figure out how to resolve the problem that in the legend always both lines are mixed up. One legend should of course show the slim black line only and the other the dashed black line. Thanks in advance!
The issue you have is that geom_vline results in a legend item that is a vertical line and geom_line gives you a horizontal line item. One solution is to create the legend kind of manually by specifying the color= aesthetic in geom_line... but not in geom_vline. You can then create a kind of "dummy" geom with geom_blank that serves as a holding object for the aesthetics of color=. You can then specify the colors for both of those items via scale_color_manual. Here's an example:
set.seed(12345)
df <- data.frame(x=1:100,y=rnorm(100))
ggplot(df, aes(x,y)) + theme_bw() +
geom_line(aes(color='score')) +
geom_vline(aes(xintercept=4), linetype=2, color='red', show.legend = FALSE) +
geom_blank(aes(color='my line')) +
scale_color_manual(name='Legend', values=c('my line'='red','score'='black'))
That creates the one legend for color... but unfortunately "my line" is solid red, when it should be dashed. To fix that, you just apply the linetype= aesthetic in the same way.
ggplot(df, aes(x,y)) + theme_bw() +
geom_line(aes(color='score', linetype='score')) +
geom_vline(aes(xintercept=4), linetype=2, color='red', show.legend = FALSE) +
geom_blank(aes(color='my line', linetype='my line')) +
scale_linetype_manual(name='Legend', values=c('my line'=2,'score'=1)) +
scale_color_manual(name='Legend', values=c('my line'='red','score'='black'))
I have two scatter plots obtained from two sets of data that I would like to overlay, when using the ggplo2 for creating single plot i am using log scale and than ordering the numbers sothe scatter plot falls into kind if horizontal S shape. Byt when i want to overlay, the information about reordering gets lost, and the plot loses its shape.
this is how the df looks like (one has 1076 entries and the other 1448)
protein Light_Dark log10
AT1G01080 1.1744852 0.06984755
AT1G01090 1.0710359 0.02980403
AT1G01100 0.4716955 -0.32633823
AT1G01320 156.6594802 2.19495668
AT1G02500 0.6406005 -0.19341276
AT1G02560 1.3381804 0.12651467
AT1G03130 0.6361147 -0.19646458
AT1G03475 0.7529015 -0.12326181
AT1G03630 0.7646064 -0.11656207
AT1G03680 0.8340107 -0.07882836
this is for single plot:
p1 <- ggplot(ratio_log_ENR4, aes(x=reorder(protein, -log10), y=log10)) +
geom_point(size = 1) +
#coord_cartesian(xlim = c(0, 1000)) +
geom_hline(yintercept=0.1, col = "red") + #check gene
geom_hline(yintercept=-0.12, col = "red") +#check gene
labs(x = "Protein")+
theme_classic()+
theme(axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank())+
labs(y = "ratio Light_Dark log10")+
labs(x="Protein")
image=p1
ggsave(file="p1_ratio_data_ENR4_cys.svg", plot=image, width=10, height=8)
and for over lay:
p1_14a <- ggplot(ratio_log_ENR1, aes(x=reorder(protein, -log10), y=log10)) +
geom_point(size = 1) +
#coord_cartesian(xlim = c(0, 1000)) +
geom_hline(yintercept=0.1, col = "red") + #check gene
geom_hline(yintercept=-0.12, col = "red") +#check gene
labs(x = "Protein")+
theme_classic()+
theme(axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank())+
labs(y = "ratio Light_Dark log10")+
labs(x="Protein")+
geom_point()+
geom_point(data=ratio_log_ENR4, color="red")
p=ggplot(ratio_log_ENR1, aes(x=reorder(protein, -log10), y=log10)) +
geom_point(size = 1) +
#coord_cartesian(xlim = c(0, 1000)) +
geom_hline(yintercept=0.1, col = "red") + #check gene
geom_hline(yintercept=-0.12, col = "red") +#check gene
labs(x = "Protein")+
theme_classic()+
theme(axis.title.x=element_blank(),
axis.text.x=element_blank(),
axis.ticks.x=element_blank())+
labs(y = "ratio Light_Dark log10")+
labs(x="Protein")
p = p + geom_point(data=ratio_log_ENR4, aes(x=reorder(protein, -log10), y=log10), color ="red" )
p
I tried to change classes... but it cant be the problem since for single plot its working like it is
The easiest solution I see for you is just binding together your two dataframes before plotting.
a$color <- 'red'
b$color <- 'blue'
ab <- a %>%
rbind(b)
ggplot(ab, aes(x = fct_reorder(protein, -log10), y = log10, color = color)) +
geom_point() +
scale_color_identity()
You can find a nice cheat-sheet for working with factors here: https://stat545.com/block029_factors.html
Following guides like ggplot Donut chart I am trying to draw small gauges, doughnuts with a label in the middle, with the intention to put them later on on a map.
If the value reaches a certain threshold I would like the fill of the doughnut to change to red. Is it possible to achieve with if_else (it would be most natural but it does not work).
library(tidyverse)
df <- tibble(ID=c("A","B"),value=c(0.7,0.5)) %>% gather(key = cat,value = val,-ID)
ggplot(df, aes(x = val, fill = cat)) + scale_fill_manual(aes,values = c("red", "yellow"))+
geom_bar(position="fill") + coord_polar(start = 0, theta="y")
ymax <- max(df$val)
ymin <- min(df$val)
p2 = ggplot(df, aes(fill=cat, y=0, ymax=1, ymin=val, xmax=4, xmin=3)) +
geom_rect(colour="black",stat = "identity") +
scale_fill_manual(values = if_else (val > 0.5, "red", "black")) +
geom_text( aes(x=0, y=0, label= scales::percent (1-val)), position = position_dodge(0.9))+
coord_polar(theta="y") +
xlim(c(0, 4)) +
theme_void() +
theme(legend.position="none") +
scale_y_reverse() + facet_wrap(facets = "ID")
Scale fill manual values= if else.... this part does not work, the error says: Error in if_else(val > 0.5, "red", "black") : object 'val' not found. Is it my error, or some other solution exists?
I also realize my code is not optimal, initially gather waited for more variables to be included in the plot, but I failed to stack one variable on top of the other. Now one variable should be enough to indicate the percentage of completion. I realise my code is redundant for the purpose. Can you help me out?
A solution for the color problem is to first create a variable in the data and then use that to map the color in the plot:
df <- tibble(ID=c("A","B"),value=c(0.7,0.5)) %>% gather(key = cat,value = val,-ID) %>%
mutate(color = if_else(val > 0.5, "red", "black"))
p2 = ggplot(df, aes(fill=color, y=0, ymax=1, ymin=val, xmax=4, xmin=3)) +
geom_rect(colour="black",stat = "identity") +
scale_fill_manual(values = c(`red` = "red", `black` = "black")) +
geom_text( aes(x=0, y=0, label= scales::percent (1-val)), position = position_dodge(0.9))+
coord_polar(theta="y") +
xlim(c(0, 4)) +
theme_void() +
theme(legend.position="none") +
scale_y_reverse() + facet_wrap(facets = "ID")
The result would be:
I have data coming from two different data frames. I am trying to create legend for each data frame. I know I can combine the data frame and do it, but because of my data source it makes the most sense to plot from two different data frames.
Please find the simplified example below. I have gotten close but the 'Main Forecast' in the legend is only white color. I want to show where 'Main Forecast' is red on the outside and white on the inside.
x = seq(1,10, 1)
y = seq(10,100, 10)
df = data.frame(x=x, y=y)
df2 = data.frame(x=5, y=50)
p = ggplot(data=df) +
geom_point(data=df,aes(x=x, y=y, color="Weekly Forecast"), fill="red", size=5, shape=16) +
geom_line(data=df,aes(x=x, y=y), color="red", size=1) +
geom_point(data=df2, aes(x=x, y=y, color="Main Forecast"), size=2, shape=16) +
scale_color_manual("Legend Title", breaks=c("Weekly Forecast", "Main Forecast"), values = c("white","red"))
p
Any assistance will be greatly appreciated.
You need to use one of the symbols that takes a fill (pch = 21:25). You then need to use override.aes to get the legend right. I've moved shared data and aes into the ggplot command.
ggplot(data=df, aes(x=x, y=y)) +
geom_point(aes(color="Weekly Forecast"), shape=16, size = 5) +
geom_line(color="red", size=1) +
geom_point(data=df2, aes(color="Main Forecast"), shape=21, fill = "white", size = 5) +
scale_color_manual("Legend Title", limits=c("Weekly Forecast", "Main Forecast"), values = c("red","red")) +
guides(colour = guide_legend(override.aes = list(pch = c(16, 21), fill = c("red", "white"))))
This can also be done without override.aes:
ggplot(data=df, aes(x=x, y=y)) +
geom_line(aes(color="Main Forecast"), size=1) +
geom_point(aes(color="Weekly Forecast", fill="Weekly Forecast"), shape=21, size = 5) +
geom_point(data=df2, aes(color="Main Forecast", fill="Main Forecast"), shape=21, size = 5) +
scale_color_manual(name="", values = c("red","red")) +
scale_fill_manual(name="", values=c("white","red"))