Merging LISS datafiles

Introduction

Goals of this script is to merge all the relevant LISS core module files into one dataframe.

Packages

#load packages
library(dplyr)
library(purrr)
library(tidyr)
library(tidyselect)
library(haven)
library(magrittr)

Data import of the different LISS modules

#create directory name
dir <- file.path("data",
          "data-processed",
          "liss_merged",
          "modules")

#create directory
if(!dir.exists(dir)) {
  dir.create(dir)
}

LISS leisure and integration module

#import the different liss files from their directory. 

#set data name
file_name <- file.path(dir,"leisure_integration_combined.Rdata")

#conditional on file.exists
if(!file.exists(file_name)) {
liss_files <- list.files("data/data-raw/leisure_integration/", full.names = T) %>%
  map(read_sav)

#For a full outer join.
liss_merged <- liss_files %>%
  reduce(full_join, by = 'nomem_encr') %>%
  arrange(nomem_encr)

#save merged data
save(liss_merged,
     file = file_name)
} else{
  load(file_name)
}

This results in a df with 14473 observations and 4909 variables.

Politics and values module

#set data name
file_name <- file.path(dir, "polval_combined.Rdata")

#conditional on file.exists
if (!file.exists(file_name)) {
  #import the data files.
  liss_files <-
    list.files("data/data-raw/pol_values_data/", full.names = T) %>%
    map(read_sav)
  
  #merge the data into one file.
  liss_pol <- liss_files %>%
    reduce(full_join, by = 'nomem_encr') %>%
    #order on nomem_encr
    arrange(nomem_encr)
  
  #save data
  save(liss_pol,
       file = file_name)
} else{
  load(file_name)
}

This results in a df with 14040 observations of 1978 variables.

#remove the household id variables from the liss_pol and liss_merged file.
# they are incomplete as in some waves they are not present. I will add them later with the background data. 
liss_pol[,c("nohouse_encr", "nohouse_encr.x", "nohouse_encr.y", "nohouse_encr.x.x", "nohouse_encr.y.y")] <- NULL
liss_merged[,c("nohouse_encr", "nohouse_encr.x", "nohouse_encr.y", "nohouse_encr.x.x", "nohouse_encr.y.y")] <- NULL

LISS CDN alter id

#set data name
file_name <- file.path(dir, "alterids_combined.Rdata")

#conditional on file.exists
if (!file.exists(file_name)) {
#import the data files.
liss_files <- list.files("data/data-raw/alter_ids/", full.names = T) %>%
  map(read_sav)

#just to be sure, order the files on nomem_encr and add suffixes to the data.
#Then we can correctly identify the different waves. 
liss_files <- lapply(1:11, function(i) {
  data <- liss_files[[i]]
  #Now I add suffixes to all the variables except the ID var, so we can correctly Identify the wave to which a variable belongs in the merged data. Otherwise we will have 11 sets of the same variables with a random .x or .y. 
  names(data)[-1] <- paste0(names(data)[-1], sep = ".", c(1:11)[i])
  return(data)
})

#merge the data
liss_alter <- liss_files %>%
  reduce(full_join, by = 'nomem_encr') %>%
  arrange(nomem_encr)

#save data
save(liss_alter, 
     file = file_name)
} else{
  load(file_name)
}

LISS background

#set data name
file_name <- file.path(dir, "background_combined.Rdata")

#conditional on file.exists
if (!file.exists(file_name)) {
#import the background variable data. 
#import data files.
liss_files <- list.files("data/data-raw/background/", full.names = T) %>%
  map(read_sav)

#just to be sure, order the files on nomem_encr and add suffixes to the data 
liss_files <- lapply(1:11, function(i) {
  data <- liss_files[[i]][base::order(liss_files[[i]]$nomem_encr),]
  names(data)[-1] <- paste0(names(data)[-1], sep = ".", c(1:11)[i])
  return(data)
})

#merge the data. 
liss_background <- liss_files %>%
  reduce(full_join, by = 'nomem_encr') %>%
  arrange(nomem_encr)

#save data
save(liss_background,
     file = file_name)
} else(
  load(file_name)
)

[1] “liss_background”

Merge all files

#merge all the data with a full outer join into one file. 
#okay, let's merge the liss_merged, liss_alter and the politics and values waves.
liss <- liss_merged %>%
  full_join(liss_alter, by = "nomem_encr") %>%
  full_join(liss_pol, by = "nomem_encr") %>%
  full_join(liss_background, by = "nomem_encr") %>%
  arrange(nomem_encr)

Full dataframe consists of 25306 observations and 7271 variables

Export data

The final DF consists of 25306 observations with 7271 variables.

#clean the global environment.
rm(list=ls()[! ls() %in% c("liss")])

save(liss, 
     file = file.path("data",
          "data-processed",
          "liss_merged",
       "liss_merged_raw.Rdata"))

---

Copyright © 2024 Jeroense Thijmen