Airbnb data adjustments.R

library(tidyverse)

import <- read_csv("original data/train_users.csv")

intro_month <- as.Date("2013-12-01")

set.seed(123)

tidy <- import %>% 
  filter(lubridate::year(date_account_created) >= 2013) %>% 
  mutate(age = if_else(age > 90 | age < 18, NA_real_, age)) %>% 
  select(id, date_account_created, gender, age, language, first_device_type, first_browser, country_destination) %>% 
  mutate(booking_val = runif(n(), 40, 200),
         booking_val = if_else(country_destination == "NDF", NA_real_, booking_val),
         friend_referral = sample(0:1, n(), prob = c(0.3, 0.7), replace = T),
         
         new_method = case_when(date_account_created < intro_month ~ 0L,
                                date_account_created <= intro_month + lubridate::dmonths(1) ~ sample(0:1, n(), prob = c(0.6, 0.4), replace = T),
                                date_account_created <= intro_month + lubridate::dmonths(2) ~ sample(0:1, n(), prob = c(0.4, 0.6), replace = T),
                                date_account_created <= intro_month -2 + lubridate::dmonths(3) ~ sample(0:1, n(), prob = c(0.2, 0.8), replace = T),
                                TRUE ~ 1L)) 


tidy %>% 
  write_csv("train_users.csv", na = "")