CODEWITHSUNDEEP
rggplot2
theamateurdataanalyst
theamateurdataanalyst

Reputation: 2834

Plotting by week with ggplot in R

I have the following data:

set.seed(123)
timeseq <- as.Date(Sys.time() + cumsum(runif(1000)*86400))
data <- rnorm(1000)
df <- data.frame(timeseq,data)

I wanted to know if anyone has any methods on how to aggregate data by week. What I am attempting to do is plot a time series ggplot, so even better if I can skip this step and have ggplot handle this. Been stuck on this all day.

Upvotes: 6

Views: 11311

Answers (4)

Kim
Kim

Reputation: 4298

I want to expand upon @chappers idea of using package lubridate, but in a fully piped way.

library(dplyr)
library(ggplot2)
library(lubridate)
set.seed(123)
data.frame(
  timeseq = as.Date(Sys.time() + cumsum(runif(1000) * 86400)),
  data = rnorm(1000)
) %>%
  mutate(timeseq = floor_date(timeseq, unit = "week")) %>%
  group_by(timeseq) %>%
  summarise(data = sum(data)) %>%
  ggplot() +
  geom_line(aes(x = timeseq, y = data))

Substitute data.frame lines with df if you have it already stored as an object.

Upvotes: 1

Peter
Peter

Reputation: 51

You can also aggregate a date aesthetic with the scale_x_date() function's breaks argument.

ggplot(df, aes(x = timeseq, y = data)) +
    stat_summary(fun.y = sum, geom = "line") +
    scale_x_date(labels = date_format("%Y-%m-%d"),
                 breaks = "1 week")

Upvotes: 3

Rorschach
Rorschach

Reputation: 32426

Another way to manually aggregate by week using dplyr.

library(dplyr)
df$weeks <- cut(df[,"timeseq"], breaks="week")
agg <- df %>% group_by(weeks) %>% summarise(agg=sum(data))
ggplot(agg, aes(as.Date(weeks), agg)) + geom_point() + scale_x_date() +
    ylab("Aggregated by Week") + xlab("Week") + geom_line()

enter image description here

Upvotes: 7

chappers
chappers

Reputation: 2415

To get the week we can use the lubridate library, with the floor_date function like so:

library(lubridate)
df$week <- floor_date(df$timeseq, "week")

We can plot the data using ggplot by doing a stats summary (there might be a better way?), and it will look like this:

stat_sum_single <- function(fun, geom="point", ...) {
  stat_summary(fun.y=fun, colour="red", geom=geom, size = 3, ...)
}

ggplot(df, aes(x=floor_date(timeseq, "week"), y=data)) + 
  stat_sum_single(sum, geom="line") + 
  xlab("week")

which will have the output:

enter image description here

Upvotes: 0

Related Questions