Nested variable by date and subject - EMA dataset - r

Question

I have created the RespNum & RespDay variables using the code below (see starting at ______________________)
Now I just need to do the following task: Create a variable called ‘Day’ that is nested by subject and date
Data sample: (click here to download)
ParticipantId DateTime_local RespNum RespDay
<chr> <dttm> <int> <int>
1 1001 2017-10-20 18:42:00 1 1
2 1001 2017-10-20 20:24:00 2 2
3 1001 2017-10-20 23:12:00 3 3
4 1001 2017-10-21 01:23:00 4 1
5 1001 2017-10-21 13:32:00 5 2
6 1001 2017-10-21 15:17:00 6 3
7 1001 2017-10-21 17:32:00 7 4
8 1001 2017-10-21 20:23:00 8 5
9 1001 2017-10-21 22:57:00 9 6
10 1001 2017-10-22 01:54:00 10 1
___________ Code used to create RespNum & RespDay ______________________
data = dataset
create new variable in correct time zone
data <- data %>%
mutate(DateTime = mdy_hm(DateTime),
DateTime_local = force_tz(DateTime, tzone = "America/New_York"))
create RespNum
this variable is the number of responses by subject.
data <- data %>%
group_by(ParticipantId) %>%
mutate(RespNum = row_number(DateTime_local)) %>%
ungroup() %>%
arrange(ParticipantId, RespNum, DateTime_local) # arrange data
data %>% select(ParticipantId, DateTime_local, RespNum) #view data
split date & time into two columns
data$date <- sapply(strsplit(as.character(data$DateTime_local), " "), "[", 1)
data$time <- sapply(strsplit(as.character(data$DateTime_local), " "), "[", 2)
change date to date format and save as numeric date
(data$date <- ymd(data$date)) #change to date format
class(data$date) #check that it is stored as date
as.numeric(data$date) #save date as numeric
class(data$date) #check that it is still date
Create RespDay Variable
ID = grouping variable
data$ID <- data$ParticipantId
date = date (not date + time)
create variable that contains subject ID and date
data$ID_DAY<-paste(data$ID,as.numeric(data$date),sep="")
data <- data %>%
group_by(ID_DAY) %>%
mutate(RespDay = row_number(date)) %>%
ungroup() %>%
arrange(ParticipantId, RespNum, RespDay, DateTime_local) # arrange data
data %>% select(ParticipantId, DateTime_local, RespNum, RespDay) #view data
The ‘Day’ variable should be a series of 1’s for the first day the participant responded, series of 2 for the 2nd day the participant responded, etc.
So using the subset of data example above:
ParticipantId DateTime_local RespNum RespDay Day
<chr> <dttm> <int> <int> <int>
1 1001 2017-10-20 18:42:00 1 1 1
2 1001 2017-10-20 20:24:00 2 2 1
3 1001 2017-10-20 23:12:00 3 3 1
4 1001 2017-10-21 01:23:00 4 1 2
5 1001 2017-10-21 13:32:00 5 2 2
6 1001 2017-10-21 15:17:00 6 3 2
7 1001 2017-10-21 17:32:00 7 4 2
8 1001 2017-10-21 20:23:00 8 5 2
9 1001 2017-10-21 22:57:00 9 6 2
10 1001 2017-10-22 01:54:00 10 1 3
Thank you!

Answer 1

Using the tidyverse and lubridate package, this works!
library(tidyverse)
library(lubridate)
##data = data name
## ParticipantId = unique subject ID
## expday = new variable created
data <- data %>%
group_by(ParticipantId) %>%
mutate(
DateTime = mdy_hm(DateTime),
Date = lubridate::date(DateTime),
expday = dense_rank(Date))
ungroup() %>%
arrange(ParticipantId, DateTime, expday) # arrange data
data %>% select(ParticipantId, DateTime, expday) #view data