forked from gmanova/excel2R
-
Notifications
You must be signed in to change notification settings - Fork 0
/
cumulative sums.R
48 lines (37 loc) · 1.55 KB
/
cumulative sums.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
library(dplyr)
library(tidyr)
getwd()
df <- read.csv("https://www.dropbox.com/s/shcok8kh2vpnaet/cumsum.csv?dl=1")
write.csv(df, "cumsum.csv")
# in the most basic form it's very easy - use the cumsum() function from base R:
df$cumsum <- cumsum(df$Sales)
df
# let's take a more realistic scenerio where you first pivot the data, and then add cumulative sum
# (cumsum) and also the total sales and % of the total achieved (so called "running total") -
# in this case there is only Date, so the cumulative runs over it
by_date <- df %>%
select(Date, Sales) %>%
group_by (Date) %>%
summarise(Sales=sum(Sales)) %>%
mutate(CumSales=cumsum(Sales)) %>%
mutate(totalSales = sum(Sales)) %>%
mutate(percent_of_total = cumsum(Sales)/sum(Sales)*100)
# in the second case, "state" is added , and the cumsum is reset for each state + date
by_date_2 <- df %>%
select(State, Date, Sales) %>%
group_by (State,Date) %>%
summarise(Sales=sum(Sales)) %>%
mutate(CumSales=cumsum(Sales)) %>%
mutate(totalSales = sum(Sales)) %>%
mutate(percent_of_total = cumsum(Sales)/sum(Sales)*100)
# in this case, we pivot rows to columns (with "spread") - and create corresponding
# cumulative sums for each
by_date_3 <- df %>%
select(State, Date, Sales) %>%
group_by (State,Date) %>%
summarise(Sales=sum(Sales)) %>%
spread(State, Sales) %>%
mutate(CumCalifornia=cumsum(California)) %>%
mutate(CumFlorida=cumsum(Florida)) %>%
mutate(CumNewYork=cumsum(NewYork)) %>%
select(Date, California, CumCalifornia,Florida,CumFlorida,NewYork,CumNewYork)