How to perform Rolling Average in R, A rolling average in time series analysis represents the average value over a predetermined number of prior periods.

Using the rollmean() function from the zoo package in R is the simplest approach to compute a rolling average:

library(dplyr)
library(zoo)

Calculate the 3-day rolling average

df %>%
mutate(rolling_avg = rollmean(values, k=3, fill=NA, align='right'))

For the column labeled values, this particular example computes a 3-day rolling average.

How to use this function is explained in the example that follows.

How to perform Rolling Average in R

Let’s say we have the following R data frame, which displays sales of a certain product over a period of 10 days:

Let’s create a data frame

df <- data.frame(day=c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10),
                 sales=c(215, 120, 114, 116, 217, 120, 112, 115, 114, 191))

df
   day sales
1    1   215
2    2   120
3    3   114
4    4   116
5    5   217
6    6   120
7    7   112
8    8   115
9    9   114
10  10   191

For each row of the data frame, we can add a new column called avg_sales3 that shows the rolling average of sales for the preceding three days:

library(dplyr)
library(zoo)

Now calculate 3-day rolling average of sales

df %>%
  mutate(avg_sales3 = rollmean(sales, k=3, fill=NA, align='right'))

    day sales avg_sales3
1    1   215         NA
2    2   120         NA
3    3   114   149.6667
4    4   116   116.6667
5    5   217   149.0000
6    6   120   151.0000
7    7   112   149.6667
8    8   115   115.6667
9    9   114   113.6667
10  10   191   140.0000

The quantity of prior periods used to generate the rolling average is controlled by the value of the k parameter in the rollmean() function.

The rolling average of the sales figure for the preceding three periods is displayed in the avg_sales3 column.

For instance, the first result of 149.6667 is determined as follows:

3-Day Moving Average is equal to 149.6667 (215 + 120 + 114) / 3.

The modify() function’s multiple rollmean() capabilities can be used to perform numerous rolling average calculations simultaneously.

The code below, for instance, demonstrates how to compute the 3-day and 4-day moving averages of sales:

ggplot2 scale in R (grammar for graphics) »