Manually adjust line colors in ggplot (geom_line)

Question

I have some data that I want to visualise using geom_line in ggplot. There are two categories in the dataset, so I use manual colors black and red for them. I also have moving average lines for each, which means that I have four lines in total.

The problem I am facing is that moving average lines are not very clear in the plot and I want to distinguish between them (using color preferrably). Currently, I can either set the same color for both moving average lines (e.g. yellow) or get the color of the original data (red and black).

What I want to do is set two new colors for the moving averages and also show them in the legend. Can you please help?

Here is an example code.

# rm(list = ls())
library(ggplot2)
X <- as.numeric(1:500)
df <- data.frame(x=X,y=rnorm(n=500,sd=.2)+sin(X*pi/110),gp=1)
for (i in 2) df <- rbind(df,data.frame(x=X,y=rnorm(n = 500,sd = .1)+sin(X*i*pi/200),gp=i))

### Classify as character ###
df$gp <- as.character(df$gp)

### Plot ###
ggplot(data = df,aes(x=x,y=y,group=gp,color=gp)) + 
  geom_line() + 
  geom_line(aes(y=rollmean(y, 15, na.pad=TRUE)), size=1, color="yellow") +
  theme_bw() +
  scale_colour_manual("", values = c("red", "black")) +
  guides(colour=guide_legend(title="Model"))

In this plot, I want the two yellow lines to be shown by two new colors. Other suggestions which can help in distinguishing between the two lines in such large datasets are also welcome.

Ronak Shah Ronak Shah · Accepted Answer · 2021-04-18T05:10:52

You can create new column with moving average, get the data in long format and then plot so that you can get legend for all the 4 colors.

library(tidyverse)
library(zoo)

df %>%
  group_by(gp) %>%
  mutate(moving_avg = rollmean(y, 15, na.pad=TRUE)) %>%
  ungroup %>%
  pivot_longer(cols = c(y, moving_avg)) %>%
  unite(col, name, gp) %>%
  ggplot(aes(x=x,y=value,color=col)) + 
  geom_line() + 
  theme_bw() +
  guides(colour=guide_legend(title="Model"))

Manually adjust line colors in ggplot (geom_line)

1 Answers