I have a dataframe with 4 columns. This includes week_no (week of the year), date (the Friday date for that week), year, sales.

I want to plot a line graph showing sales for the winter period by week. The winter period runs from week 50 - week 4. Each winter season should it’s own line and colour.

Sample dataframe code

library(ggplot2)

df <- data.frame(
week_no = c("50","51","52","1","2","3","4","50","51","52","1","2","3","4"),
date = c("2018-12-14", "2018-12-21", "2018-12-28", "2019-01-04", "2019-01-11", "2019-01-18", "2019-01-25", "2019-12-13", "2019-12-20", "2019-12-27", "2020-01-03", "2020-01-10", "2020-01-17", "2020-01-24"), 
         year = c("2018", "2018", "2018", "2019", "2019", "2019", "2019", "2019", "2019", "2019", "2020", "2020", "2020", "2020"),
sales = c(546,873,532,424,235,321,531,865,869,458,234,654,345,984))

Sample dataframe

I’ve tried the below code and it gives a result close to what I need other than the but the years don’t match up correct and the week order is continuous running 1-52.

I desired the x-axis to be ordered as 50,51,52,1,2,3 and each line to show data similarly by date order rather than year.

Code

winter_line_plot <- ggplot(df, aes(x = week_no, y = sales))   
  geom_line(aes(color = year))

I tried using the dates for x-axis but my result showed one continuous line rather than separate colour/lines by the different winter seasons.

CodePudding user response：

You need to add year as an aesthetic. Keep week discrete, and add scale limits. In order to get the season, your idea in the comment to create a grouping variable for each season is certainly the most straight forward.

library(tidyverse)

df <- data.frame(
  week_no = c("50","51","52","1","2","3","4","50","51","52","1","2","3","4"),
  date = c("2018-12-14", "2018-12-21", "2018-12-28", "2018-01-04", "2018-01-11", "2018-01-18", "2018-01-25", "2019-12-13", "2019-12-20", "2019-12-27", "2020-01-03", "2020-01-10", "2020-01-17", "2020-01-24"), 
  year = c("2018", "2018", "2018", "2019", "2019", "2019", "2019", "2019", "2019", "2019", "2020", "2020", "2020", "2020"),
  sales = c(546,873,532,424,235,321,531,865,869,458,234,654,345,984))

##  if your data is indeed very regular and contains always one and no more than one row per week,
## then you can easily create the groups with rep
## (make sure the data frame is sorted by year and week)
df %>% 
  arrange(year, week_no) %>%
  mutate(season = rep(c("winter_1819"," winter_1920"), each = 7)) %>%
ggplot()  
  ## week is discrete, so you need to specify group
  geom_line(aes(week_no, sales, color = season, group = season))  
  ## set axis breaks
  scale_x_discrete(limits = as.character(c(50:52, 1:4) ))