Restructuring data depending on reoccuring values in R - r

Question

Right now, I try to restructure my data (it's about responsiveness of contacted people in a survey) which has a structure like this:
df_test <- data.frame(
Residence=c(rep("Berlin",10),rep("Frankfurt",10),rep("Munich",10)),
Response=c(rep(TRUE,14),rep(FALSE,16)),
ID=c(rep(1:15,each=2)),
Contact = c(rep(c("Phone","Mail","In_Person","Phone","eMail","Phone"))),
Date = sample(seq(as.Date('2000/01/01'), as.Date('2001/01/01'), by="day"), 30)
)
df_test <- df_test[order(df_test$ID,df_test$Date),]
In the resulting dataframe, each line represents one contact event and, usually, all people (labelled by ID) have been contacted multiple times by different means:
#first 4 lines of dataframe:
Residence Response ID Contact Date
2 Berlin TRUE 1 Mail 2000-07-25
1 Berlin TRUE 1 Phone 2000-09-25
3 Berlin TRUE 2 In_Person 2000-02-06
4 Berlin TRUE 2 Phone 2000-10-01
To get a nice overview with focus on the contacted people for e.g. plots, I want to create a new data frame in which every line represents one contacted person, with fixed values just appearing once (e.g. ID, Residence, Response) while contact-specific values (Contact, Date) are listed in each line like so:
#restructured lines in new dataframe from first 4 lines of original dataframe:
Residence Response ID Contact Date Contact.1 Date.1
1 Berlin TRUE 1 Mail 2000-07-25 Phone 2000-09-25
2 Berlin TRUE 2 In_Person 2000-02-06 Phone 2000-10-01
With the initial sorting by date i hope to also get the contact attempts in each line in chronological order.
While i don't have any code which is close to running, i tried to at least get a dataframe with an empty column and fill it with with the extracted IDs, without duplicates:
for (i in df_test[,"ID"]){
if (df_test[i,"ID"] != df_test [i-1,"ID"]){
df_test_restructured<-append(df_test_restructured,df_test[i,"ID"])
}
}
After many unfruitful attmempts, I figured there should be some existing and more efficient strategies or functions unknown to me. Any suggestions? Thanks in advance <3
EDIT: Ideally, each row would have the contact attempt listed in order, since people also have been contacted multiple times with the same medium. I want to extract info like e.g. people have mostly responded after the first reminder email after already having been sent an initial email

Answer 1

Assuming you want per person (ID) one row to show at what date with what (phone, email,..) there was a contact you could do something like this with tidyverse.
library(tidyverse)
df_test <- data.frame(
Residence=c(rep("Berlin",10),rep("Frankfurt",10),rep("Munich",10)),
Response=c(rep(TRUE,14),rep(FALSE,16)),
ID=c(rep(1:15,each=2)),
Contact = c(rep(c("Phone","Mail","In_Person","Phone","eMail","Phone"))),
Date = sample(seq(as.Date('2000/01/01'), as.Date('2001/01/01'), by="day"), 30)
)
df_test %>%
group_by(ID) %>%
pivot_wider(names_from = Contact, values_from = Date)
#> # A tibble: 15 x 7
#> # Groups: ID [15]
#> Residence Response ID Phone Mail In_Person eMail
#> <chr> <lgl> <int> <date> <date> <date> <date>
#> 1 Berlin TRUE 1 2000-01-04 2000-09-06 NA NA
#> 2 Berlin TRUE 2 2000-03-15 NA 2000-05-19 NA
#> 3 Berlin TRUE 3 2000-11-05 NA NA 2000-05-06
#> 4 Berlin TRUE 4 2000-11-02 2000-03-29 NA NA
#> 5 Berlin TRUE 5 2000-12-20 NA 2000-04-30 NA
#> 6 Frankfurt TRUE 6 2000-02-23 NA NA 2000-02-05
#> 7 Frankfurt TRUE 7 2000-08-30 2000-11-29 NA NA
#> 8 Frankfurt FALSE 8 2000-02-20 NA 2000-08-08 NA
#> 9 Frankfurt FALSE 9 2000-12-11 NA NA 2000-05-25
#> 10 Frankfurt FALSE 10 2000-12-21 2000-01-15 NA NA
#> 11 Munich FALSE 11 2000-07-07 NA 2000-12-16 NA
#> 12 Munich FALSE 12 2000-08-26 NA NA 2000-09-03
#> 13 Munich FALSE 13 2000-05-02 2000-11-20 NA NA
#> 14 Munich FALSE 14 2000-04-05 NA 2000-09-30 NA
#> 15 Munich FALSE 15 2000-09-26 NA NA 2000-05-22
New Addition based on your new target
I am not sure if this is the tidiest way, but I guess it is what you are looking for.
df_test %>%
group_by(ID) %>%
arrange(Date) %>%
mutate(no = row_number()) %>%
pivot_wider(names_from = c(no), values_from = c(Contact,Date)) %>%
select(c(Residence:Contact_1, Date_1, Contact_2, Date_2)) %>%
arrange(ID)
#> # A tibble: 15 x 7
#> # Groups: ID [15]
#> Residence Response ID Contact_1 Date_1 Contact_2 Date_2
#> <chr> <lgl> <int> <chr> <date> <chr> <date>
#> 1 Berlin TRUE 1 Mail 2000-01-09 Phone 2000-04-26
#> 2 Berlin TRUE 2 Phone 2000-01-27 In_Person 2000-10-14
#> 3 Berlin TRUE 3 eMail 2000-03-01 Phone 2000-07-14
#> 4 Berlin TRUE 4 Phone 2000-05-19 Mail 2000-09-22
#> 5 Berlin TRUE 5 Phone 2000-07-06 In_Person 2000-12-03
#> 6 Frankfurt TRUE 6 eMail 2000-07-05 Phone 2000-11-20
#> 7 Frankfurt TRUE 7 Phone 2000-02-06 Mail 2000-12-28
#> 8 Frankfurt FALSE 8 Phone 2000-04-03 In_Person 2000-09-06
#> 9 Frankfurt FALSE 9 eMail 2000-06-16 Phone 2000-06-24
#> 10 Frankfurt FALSE 10 Phone 2000-01-26 Mail 2000-05-02
#> 11 Munich FALSE 11 In_Person 2000-02-15 Phone 2000-06-28
#> 12 Munich FALSE 12 eMail 2000-03-22 Phone 2000-04-24
#> 13 Munich FALSE 13 Phone 2000-03-21 Mail 2000-08-02
#> 14 Munich FALSE 14 In_Person 2000-09-01 Phone 2000-11-27
#> 15 Munich FALSE 15 Phone 2000-05-27 eMail 2000-07-09

Answer 2

You can start by doing:
> df_test %>%
+ pivot_wider(names_from = Contact,values_from=Date)
# A tibble: 15 x 7
Residence Response ID Phone Mail In_Person eMail
<fct> <lgl> <int> <date> <date> <date> <date>
1 Berlin TRUE 1 2000-01-20 2000-02-18 NA NA
2 Berlin TRUE 2 2000-07-24 NA 2000-03-19 NA
Actually, plotting with your original df is really doable.