Plotting “time of day” data using ggplot2

April 14, 2010
By

(This article was first published on What You're Doing Is Rather Desperate » R, and kindly contributed to R-bloggers)

William asks:

How can I make a graph that looks like this, “tweet density” style, showing time intervals?

He then helpfully describes his input data: a CSV file with headers “time started, time finished, date”.

Here’s a simple CSV file, tasks.csv:

task,date,start,end
task1,2010-03-05,09:00:00,13:00:00
task2,2010-03-06,10:00:00,15:00:00
task3,2010-03-06,11:00:00,18:00:00
task4,2010-03-07,08:00:00,11:00:00
task5,2010-03-08,14:00:00,17:00:00
task6,2010-03-09,12:00:00,16:00:00
task7,2010-03-10,14:00:00,19:00:00
task8,2010-03-11,09:30:00,13:30:00

Read into R, calculate the weekday and reorder the weekday factors from Sunday, Monday…to Saturday:

tasks <- read.csv("tasks.csv", header = T)
# day of week
tasks$day <- weekdays(strptime(tasks$date, "%Y-%m-%d"))
week      <- c("Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday")
tasks$day <- factor(tasks$day, levels = week)

Convert the start and end times to decimal hours. I’m not very familiar with the as.POSIX… functions, so I’m sure that there’s a more elegant way to do this:

# convert time to decimal hours
tasks$start.ct   <- as.POSIXct(paste(tasks$date, tasks$start, sep = " "))
tasks$end.ct     <- as.POSIXct(paste(tasks$date, tasks$end, sep = " "))
tasks$start.hour <- as.POSIXlt(tasks$start.ct)$hour + as.POSIXlt(tasks$start.ct)$min/60 + as.POSIXlt(tasks$start.ct)$sec/3600
tasks$end.hour   <- as.POSIXlt(tasks$end.ct)$hour + as.POSIXlt(tasks$end.ct)$min/60 + as.POSIXlt(tasks$end.ct)$sec/3600

We’re going to plot task duration as a horizontal rectangle. If there is more than one task per day, we need to offset the rectangle vertically, so as they don’t overlap.

# offset tasks if > 1 per day
tasks$ymin <- c(rep(0, nrow(tasks)))
t <- table(tasks$day)
for(day in rownames(t)) {
  if(t[[day]] > 1) {
    ss <- tasks[tasks$day == day,]
    y  <- 0
    for(i in as.numeric(rownames(ss))) {
      tasks[i,]$ymin <- y
      y <- y + 1.2
    }
  }
}

Finally, call ggplot with the rectangle geom plus a bunch of options to colour the rectangles (by task), facet the plot (by day) and clean up, rescale and label the axes:

# plot
library(ggplot2)
png(filename = "tasks.png", width = 640, height = 480)
ggplot(tasks, aes(xmin = start.hour, xmax = end.hour, ymin = ymin, ymax = ymin + 1, fill = factor(task))) + geom_rect() +
facet_grid(day~.) + opts(axis.text.y = theme_blank(), axis.ticks = theme_blank()) + xlim(0,23) + xlab("time of day")
dev.off()

Result:


Filed under: programming, R, statistics Tagged: datetime, how to, plotting

To leave a comment for the author, please follow the link and comment on his blog: What You're Doing Is Rather Desperate » R.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Tags: , , , , ,

Comments are closed.