R - plotting multiple time series, same x-axis values, but time is mixed within data frame

Question

I am trying to plot the following data frame, where there are 3 different time series (identified by user0, user1, and user2). Each row has a user identifier, the date, and a value.

> df
   userId       date steps
1   user0 2016-03-24   794
2   user0 2016-03-25   562
3   user0 2016-03-26   682
4   user0 2016-03-27   722
5   user0 2016-03-28   883
6   user1 2016-03-24  3642
7   user1 2016-03-25  3776
8   user1 2016-03-26  3585
9   user1 2016-03-27  3585
10  user1 2016-03-28  3471
11  user2 2016-03-24  5959
12  user2 2016-03-25  5933
13  user2 2016-03-26  5802
14  user2 2016-03-27  6094
15  user2 2016-03-28  5903
> dput(df)
structure(list(userId = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 
2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L), .Label = c("user0", "user1", 
"user2"), class = "factor"), date = structure(c(16884, 16885, 
16886, 16887, 16888, 16884, 16885, 16886, 16887, 16888, 16884, 
16885, 16886, 16887, 16888), class = "Date"), steps = c(794L, 
562L, 682L, 722L, 883L, 3642L, 3776L, 3585L, 3585L, 3471L, 5959L, 
5933L, 5802L, 6094L, 5903L)), .Names = c("userId", "date", "steps"
), row.names = c(NA, -15L), class = "data.frame")

I would like plot all the time series (however many there are identified by the userId field) using different colors and with the date as the x-axis . I tried the following, but as you can see, the dates are repeated on the x-axis.

plot(df$steps, axes=F, xlab="", ylab="Steps", ylim=c(0,max(df$steps)))
axis(2)
axis(1, at = seq_along(df$date), labels = df$date, las = 2, cex.axis = 0.70)
box()

I looked at other postings, such as " Plot multiple lines (data series) each with unique color in R " and " Plotting multiple time series on the same plot using ggplot() ", but they do not have my problem of the time variable being mixed in with the other data.

A solution using color lines with and without ggplot would be greatly appreciated.

Answer 1

Here is a base R version:

plot(0, 0, type = "n", xlim = range(df$date), ylim = c(0, max(df$step)), axes = FALSE, xlab = "", ylab = "steps")
axis(2, las = 1)
axis(1, at = df$date, labels = df$date, las = 2, cex.axis = 0.70)
box()

cols <- c("red", "green", "blue")
for (i in 1:length(unique(df$userId)))
  with(df[df$userId == unique(df$userId)[i], ], lines(date, steps, col = cols[i]))

Answer 2

With ggplot:

library(ggplot2)
ggplot(df, aes(x = date, y = steps, colour = userId)) + geom_line()

---

An equivalent (but still pretty ugly) base R version takes a lot more work:

plot(0, type = 'n', axes = FALSE, xlab = 'date', ylab = 'steps',
     xlim = c(min(df$date), max(df$date)), 
     ylim = c(min(df$steps) - 100, max(df$steps) + 100))
axis.Date(1, df$date, format = '%F')    # `axis.Date` is helpful here
axis(2, seq(0, max(df$steps + 500), 500))
box()
lapply(split(df, df$userId), function(x){lines(x$date, x$steps, 
                                               col = as.numeric(substr(x$userId, 5, 5)) + 1)})
# `paste` extra space to align legend correctly...*sigh*
legend('bottomright', paste(levels(df$userId), '   '), col = 1:3, lty = 1)

Note that it needs a good bit of fine-tuning.