Plotting of cumulative data of multiple year in R

Question

I have daily rainfall data which I have converted to yearwise cumulative value using following code

library(seas)
library(data.table)
library(ggplot2)

#Loading data
data(mscdata)
dat <- (mksub(mscdata, id=1108447))
dat$julian.date <- as.numeric(format(dat$date, "%j"))
DT <- data.table(dat)
DT[, Cum.Sum := cumsum(rain), by=list(year)]

df <- cbind.data.frame(day=dat$julian.date,cumulative=DT$Cum.Sum)

But when I am trying to plot, it gives me weired output like

#Plotting using base R
df <- df[order(df[,1]),]
plot(df$day, df$cumulative, type="l", xlab="Day", ylab="Cumulative rainfall")

Same I am getting using ggplot2 also like

#Plotting using ggplot2
ggplot(df, aes(x = day, y = cumulative)) + geom_line()

But I want to get lines for every year may be in grey colour and mean value over the years in red colour like the following figure

How to achieve this?

It worked thank you very much. For base R plotting what can be done? — UseR10085, Jun 25 '20 at 10:51
Plot all as you did, but add `type="n"` and then use `lines` within each year by splitting on year, using `lapply` or `tapply` or `split` or ... — Edward, Jun 25 '20 at 11:02

score 1 · Answer 1 · edited Jun 25 '20 at 11:36

So, as you can see, you lost the group in geom_line. Without group, ggplot will connect all dots sharing with the axis.x. Here is an example to add 'year' as the group and calculate the mean value for each day.

library(reshape2)

data(mscdata)
dat <- (mksub(mscdata, id=1108447))
dat$julian.date <- as.numeric(format(dat$date, "%j"))
DT <- data.table(dat)
DT[, Cum.Sum := cumsum(rain), by=list(year)]

dt <- cbind.data.frame(day=dat$julian.date,cumulative=DT$Cum.Sum,year=DT$year)
TB <- melt(dt, id.vars = c('day','year'))
Mean_l = colMeans(reshape(TB[c("day",'year','value')],timevar='day',idvar = 'year', direction = 'wide'),na.rm = T)
Mean_l= Mean_l[-1]
Mean_l <- data.frame(day=c(1:length(Mean_l)),Mean_l)

TB_f <- data.frame(TB,avr=Mean_l$Mean_l[match(TB$day,Mean_l$day)])

ggplot(TB_f,aes(day,value))+ geom_line(aes(group=year))+ geom_line(aes(y=avr),color='red')+ theme_light()

Thank you very much. Your answer works but it appears a bit messy which can be reduced using `tidyverse` package. — UseR10085, Jun 25 '20 at 11:38

Edward · Accepted Answer · 2020-06-25T12:23:03.217

Add the group aesthetic to tell ggplot to group by the year and add stat_summary to draw the red line (without the grouping).

library(ggplot2)

ggplot(DT, aes(x = julian.date, y = Cum.Sum, group=year)) + 
  geom_line(col="grey") +
  labs(x="Date", y="Cumulative sum") +
  stat_summary(aes(group=NULL), fun="mean", geom="line", col="red", se="none", lwd=1)

For base graphics, it's a little more involved:

par(mar=c(4,3.5,1,1))
plot(df$day, df$cumulative, type="n", xlab="Day", ylab="Cumulative rainfall", las=1)
grid()

lapply(split(df, df$year), FUN=function(x) 
     with(x, lines(day, cumulative, col="grey", lwd=0.5)))

with(aggregate(cumulative~day, FUN=mean, data=df), 
    lines(x=day, y=cumulative, lwd=2, col="red"))

Excellent answer... Please visit [this](https://stackoverflow.com/questions/62574268/applying-yearwise-segmented-regression-in-r), I think you will be able to answer that question also. — UseR10085, Jun 25 '20 at 12:16

Plotting of cumulative data of multiple year in R

2 Answers2