Lag or lead to assess user paths in R

Question

I have the following dataset:

User	Session_ID	Page	Path_Number
123A	12345	home	1
123A	12345	services	2
123A	12345	pricing	3
123A	12345	about	4
123A	12345	services	5
123A	12345	home	6
123B	34567	home	1
123B	34567	services	2
123B	34567	about	3
123B	34567	multimedia	4
123C	56789	home	1
123C	56789	about	2
123C	56789	pricing	3
123C	56789	about	4
123C	56789	services	5

There are three users with unique session IDs. Path Number is the path they follow once they are on the website. And, Page is the pages they visit.

The question that I am trying to answer is: How many people first go to the 'services' page and then go to the 'about' page?

I am using the following code to assess which user and session have both 'services' and 'about' in the path:

    dataset %>% group_by(Session_ID, User) %>% 
      summarize(services_and_about = ('services' %in% Page) & ('about' %in% Page)) %>%
      filter(services_and_about == "TRUE")

The result would be users 123A, 123B, and 123C.

However, I would like to also know which users visit the 'services' page BEFORE the 'about' page (only users 123A and 123B). I know I should use a lag or lead function here, but I am not sure how.

Thanks a lot for helping!

Answer 1

Try this:

library(dplyr)
dat %>%
  group_by(Session_ID, User) %>%
  summarize(
    services_before_about = all(c("services", "about") %in% Page) &
      any(Path_Number[Page == "services"] < Path_Number[Page == "about"]),
    .groups = "drop")
# # A tibble: 3 x 3
#   Session_ID User  services_before_about
#        <int> <chr> <lgl>                
# 1      12345 123A  TRUE                 
# 2      34567 123B  TRUE                 
# 3      56789 123C  FALSE                

Answer 2

Assuming that other values of Page are irrlevant for this question, you can first filter the data to services and about only, then use a lag function to see whether there is any occurence of about that is preceded by services:

dataset %>% 
  filter(Page %in% c("services", "about")) %>%
  group_by(Session_ID, User) %>%
  mutate(previous_page = dplyr::(Page)) %>%
  mutate(about_after_service = ifelse(
    Page == "about" & previous_page == "services", 1, 0)) %>%
  group_by(User) %>% 
  summarize(about_after_service_n = sum(about_after_service)) %>%
  filter(about_after_service_n > 0)