CODEWITHSUNDEEP
rggplot2time-seriesheatmap
domwoe
domwoe

Reputation: 119

Plot timeseries as heatmap

I have a data frame with timestamps and values (a month with values approximately every 5min). Now I want to have plot with the day on the x-axis, the time of day on the y-axis and the value as the color. Best would be if I can control the binning on the y-axis (5min, 10min, 15min.. with the value as the mean for each bin).

I am experimenting with ggplot2, but I can't get a reasonable result.

p_heat <- ggplot(data = data,
                 aes(x = days(timestamp),
                     y = minutes(timestamp) + hours(timestamp) * 60, fill = value)) +  
            geom_tile()

Test data can be generated as follows:

data <- data.frame(timestamp = seq(from = as.POSIXct("2013-09-01 00:00:00"), 
                                   to = as.POSIXct("2013-10-01 00:00:00"), by = "5 mins"),
                   value = runif(8641, 0, 1))

This was the best I could get so far :) I also tried playing with scale_x/y_date()

I would be glad about some hints which push me into the right direction. Thanks!

Upvotes: 4

Views: 7072

Answers (2)

Henrik
Henrik

Reputation: 67778

Is it something like this you are looking for?

library(ggplot2)   

# create date variable for the x-axis
df$date <- as.Date(df$timestamp, format = "%Y-%m-%d")

# get H:M components
df$hm <- format(df$timestamp, "%H:%M")    

# create y-axis breaks and labels
lab <- with(df, paste(format(df$timestamp, "%H"), "00", sep = ":"))

gg <- ggplot(data = df, aes(x = date, y = hm, fill = value)) +
  geom_tile() +
  scale_y_discrete(breaks = lab)

gg

enter image description here

Upvotes: 8

keithing
keithing

Reputation: 281

Perhaps try this:

p_heat <- ggplot(data=data, aes(x=timestamp$mday ,y=timestamp$min+timestamp$hour*60)) 
    + stat_density2d(geom="tile", aes(fill = value), contour = FALSE)

Here's the relevent documentation: http://docs.ggplot2.org/current/stat_density2d.html

Upvotes: 0

Related Questions