R: Moving average length of time between two dates - r

Question

I have a dataset of observations with start and end dates. I would like to calculate the moving average difference between the start and end dates.
I've included an example dataset below.
require(dplyr)
df <- data.frame(id=c(1,2,3),
start=c("2019-01-01","2019-01-10", "2019-01-05"),
end=c("2019-02-01", "2019-01-15", "2019-01-10"))
df[,c("start", "end")] <- lapply(df[,c("start", "end")], as.Date)
id start end
1 2019-01-01 2019-02-01
2 2019-01-10 2019-01-15
3 2019-01-05 2019-01-10
The overall date ranges are 2019-01-01 to 2019-02-01. I would like to calculate the average difference between the start and end dates for each of the dates in that range.
The result would look exactly like this. I've included the actual values for the averages that should show up:
date avg
2019-01-01 0
2019-01-02 1
2019-01-03 2
2019-01-04 3
2019-01-05 4
2019-01-06 3
2019-01-07 4
2019-01-08 5
2019-01-09 6
2019-01-10 7
2019-01-11 5.5
. .
. .
. .

Answer 1

Creating a reproducible example:
df <- data.frame(id=c(1,2,3,4),
start=c("2019-01-01","2019-01-01", "2019-01-10", "2019-01-05"),
end=c("2019-01-04", "2019-01-05", "2019-01-12", "2019-01-08"))
df[,c("start", "end")] <- lapply(df[,c("start", "end")], as.Date)
df
Returns:
id start end
1 2019-01-01 2019-01-04
2 2019-01-01 2019-01-05
3 2019-01-10 2019-01-12
4 2019-01-05 2019-01-08
Then using the group_by function from dplyr:
library(dplyr)
df %>%
group_by(start) %>%
summarize(avg=mean(end - start)) %>%
rename(date=start)
Returns:
date avg
<time> <time>
2019-01-01 3.5 days
2019-01-05 3.0 days
2019-01-10 2.0 days

Answer 2

Editing the answer as per comments.
Creating the df:
require(dplyr)
df <- data.frame(id=c(1,2,3),
start=c("2019-01-01", "2019-01-10", "2019-01-05"),
end=c("2019-02-01", "2019-01-15", "2019-01-10"))
df[,c("start", "end")] <- lapply(df[,c("start", "end")], as.Date)
Create dates for every start-end combination:
#gives the list of all dates within start and end frames and calculates difference
datesList = lapply(1:nrow(df),function(i){
dat = data_frame('date'=seq.Date(from=df$start[i],to=df$end[i],by=1),
'start'=df$start[i]) %>%
dplyr::mutate(diff=date-start)
})
Finally, group_by the date and find avg to give output exactly as the one in the question:
finalDf = bind_rows(datesList) %>%
dplyr::filter(diff != 0) %>%
dplyr::group_by(date) %>%
dplyr::summarise(avg=mean(diff,na.rm=T))
The output thus becomes:
# A tibble: 31 x 2
date avg
<date> <time>
1 2019-01-02 1.0 days
2 2019-01-03 2.0 days
3 2019-01-04 3.0 days
4 2019-01-05 4.0 days
5 2019-01-06 3.0 days
6 2019-01-07 4.0 days
7 2019-01-08 5.0 days
8 2019-01-09 6.0 days
9 2019-01-10 7.0 days
10 2019-01-11 5.5 days
# … with 21 more rows
Let me know if it works.