Convert multiple cumulative columns to normal form and plot in R-CodePudding

So I have a data frame of the following form represented in a very simple way:

Sno	Box1	Box2
1	5	8
2	7	9
3	13	12

But it consists of 50 such boxes and 100 such serial numbers

Now I am facing two problems:

Firstly, I want to convert this cumulative data to normal data

Sno	Box1	Box2
1	5	8
2	2	1
3	6	3

Secondly, I want to plot each of these columns in a graph. So the plot 1 will produce a graph for Box 1 with the SNo as the x-axis and the box 1 values as the y-axis. Similarly, plot 2 will produce a graph for Box 2 with the SNo as the x-axis and the box 2 values as the y-axis, and so on.

So the end result would be 50 graphs represented in a paper, which will show the data in a daily basis, using R.

CodePudding user response：

Your data is (please provide code, not tables/ pictures)

df <- data.frame(Sno= 1:3, Box1= c(5,7,13), Box2= c(8,9,12))

We can simply use difference between rows to make it "normal" and not cumulative:

df <- cbind.data.frame(Sno= df$Sno,
                       rbind.data.frame(df[1, 2:ncol(df)],
                                        df[2:(nrow(df)), 2:ncol(df)] - 
                                        df[1:(nrow(df)-1), 2:ncol(df)])
                                        )

And you can plot all the columns like that:

for(col_i in names(df)[-1]){
  plot(df$Sno, df[ , col_i], ylab= col_i)
}