Suggestions for examples

[MarkEdmondson1234]

I asked on Twitter/Slack for examples, here were some suggestions:

• Sequence segments
• For the top conversion paths, seeing the distribution of positions in the path (start, mid, end) for each channel/source/medium/campaign to get an idea of where each channel/source/medium/campaign is most likely to appear in the conversion path
• Visualizing user journey based on content grouping, we can use history converted users.
• Predicting next page and chances to convert, where do we want the user to go next (maybe pop-ups?)
• Sending data from GA to a client by using whatsapp chat. (group chat)
• Another one could be visualizing URLs visited by Client ID, orderd by Hit Timestamp. I think it could be a really fast, neat, and accurate way of watching users movement through the site. Could also be done with User IDs for Cross Device data
• pulling mfc report data by campaign dimension, and replacing the unknown with direct or organic accordingly
• merging the same data but from dimension source/medium with dimension campaign. Any of those would be really nice.
• Pulling multi touch data with the "event" timestamps in order to calculate the predicted conversion time. Everyone focuses on the chance it occurs, I'd like to predict when...

[MarkEdmondson1234]

This could be a basic example - write a group of GA fetches to one CSV file:

## setup
library(googleAnalyticsR)
library(tidyverse)

# names the auth file so you can reuse it later
ga_auth("ga.auth")

viewIds <- c(12345678, 23456789, 34567891, 45678912)

# function to map over the viewID data fetches
get_ga_data <- function(viewId){
  google_analytics(viewId,
                   #yyyy-mm-dd
                   date_range = c("2017-11-01","2018-10-31"), 
                   metrics = c("metric1"), 
                   dimensions = c("dimension1","dimension2"),
                   anti_sample = TRUE
  )
}

list_of_gadata <- map(viewIds, get_ga_data)
# name the list of data.frames so you know which viewId they are
list_of_gadata <- setNames(list_of_gadata, viewIds)

# function to process each data.frame
process_ga <- function(x){
  x %>% 
    mutate(transactionRevenue = if(is.null(transactionRevenue)) 0 else transactionRevenue,
           transactionId = if(is.null(transactionId)) "NA" else transactionId,
           dimension01 = if(is.null(dimension01)) "NA" else dimension01
           )
}

## process all the data.frames
list_of_gadata_processed <- map(list_of_gadata, process_ga)

name_to_col <- function(the_name, the_obj){
  # get the right object from the list
  o <- the_obj[[the_name]]

  o %>% 
    mutate(viewId = the_name)
}

# could do this in a lookup if a lot of them, otherwise in same order as viewIds
view_names <- c("View1","View2","View3","View4")

## put the viewId into a column of the data.frames
finished_list <- map(viewIds, ~name_to_col(.x, the_obj = list_of_gadata_processed))

# combine data.frames and write to csv
all_data <- finished_list %>% 
  reduce(bind_rows) %>% 
  write_csv("data/place_to_save.csv")