Overlapping legend of a line plot

Question

I want to show each line as having a different color and shape, and that should be reflected on the legend. Basically, it should look something like this My Goal.

I could make the color and point style unique with the following code.

data <- data.frame("1" = c(1, 3, 6, 4), "2" = c(2, 9, 2, 8))
colnames(data) <- paste0("", c(1,2))

power_1 <- factor(c(0,0.25,0.33,"Period"))

density_vs_power <- ggplot(data = as.data.frame(data))


# Loop to add geom_line layers for each column
for (i in 1:2) {
  density_vs_power <- density_vs_power +
    geom_line(aes(x = power_1, y = !!sym(as.character(i)), group = 1, color = as.character(i))) +
    geom_point(shape = i, aes(x = power_1, y = !!sym(as.character(i)), color = as.character(i)))
}

density_vs_power <- density_vs_power +
  scale_y_continuous(trans = "log10") +
  theme_classic()


density_vs_power

However, this does not show a legend, as shown here different color but not legend. I knew that if I put the color attribute into aes(), a legend should appear. When I attempt this with the following code,

data <- data.frame("1" = c(1, 3, 6, 4), "2" = c(2, 9, 2, 8))
colnames(data) <- paste0("", c(1,2))

power_1 <- factor(c(0,0.25,0.33,"Period"))

density_vs_power <- ggplot(data = as.data.frame(data))


# Loop to add geom_line layers for each column
for (i in 1:2) {
  density_vs_power <- density_vs_power +
    geom_line(aes(x = power_1, y = !!sym(as.character(i)), group = 1, color = as.character(i))) +
    geom_point(shape = i, aes(x = power_1, y = !!sym(as.character(i)), color = as.character(i)))
}

density_vs_power <- density_vs_power +
  scale_y_continuous(trans = "log10") +
  theme_classic()


density_vs_power

the legend overlaps, as shown here legend but overlapping features. How can make the legend separate for each line and basically show the format "Object #"?

Any help is much appreciated!

Answer 1

The easiest way to achieve your desired result would be to reshape your data to long or tidy format which also allows to get rid of the for loop:

library(ggplot2)

data_long <- data |>
  dplyr::mutate(power1 = power_1) |>
  tidyr::pivot_longer(
    -power1,
    names_to = "group"
  )

ggplot(
  data = data_long,
  aes(power1, value, color = factor(group), group = group)
) +
  geom_line() +
  geom_point(aes(shape = factor(group))) +
  scale_y_continuous(trans = "log10") +
  theme_classic()

Answer 2

Is this what you mean?

library(tidyverse)
my_data <- data.frame("One" = c(1, 3, 6, 4), "Two" = c(2, 9, 2, 8), power_1=factor(c(0,0.25,0.33,"Period")))
# avoid using numerical numbers in column names.
# and a data.frame can hold all kinds of data together

# make the data long
my_data <- my_data %>% pivot_longer(cols=c("One", "Two"), names_to = "level", values_to = "value")

#plot it!
ggplot(data = my_data, aes(x=power_1, y=value, group = level, col=level)) + 
  geom_line() +
  geom_point(aes(pch=level)) +
  theme_bw()  # I like the bw theme..