Initialize repository

03_data/02_anonymized_data/cleaning.R

@@ -0,0 +1,117 @@
+#' ---
+#' title: "HMC data cleaning"
+#' author: "Nora Wickelmaier"
+#' date: 2025-10-14
+#' ---
+
+# clear workspace
+rm(list = ls())
+
+# Read anonymized data
+dat1 <- read.csv("HMC_wave1_anonymized.csv", na.strings = c("-99", "NA"))
+dat2 <- read.csv("HMC_wave2_anonymized.csv", na.strings = c("-99", "NA"))
+dat3 <- read.csv("HMC_wave3_anonymized.csv", na.strings = c("-99", "NA"))
+dat4 <- read.csv("HMC_wave4_anonymized.csv", na.strings = c("-99", "NA"))
+dat5 <- read.csv("HMC_wave5_anonymized.csv", na.strings = c("-99", "NA"))
+dat6 <- read.csv("HMC_wave6_anonymized.csv", na.strings = c("-99", "NA"))
+
+# Check if subjects are double within waves
+which(table(dat1$subj_id) > 1)
+which(table(dat2$subj_id) > 1)
+which(table(dat3$subj_id) > 1)
+which(table(dat4$subj_id) > 1)
+which(table(dat5$subj_id) > 1)
+which(table(dat6$subj_id) > 1)
+
+dat1[dat1$subj_id == names(which(table(dat1$subj_id) > 1)),
+     c("StartDate", "subj_id", "Duration_in_seconds")]
+dat3[dat3$subj_id == names(which(table(dat3$subj_id) > 1)),
+     c("StartDate", "subj_id", "Duration_in_seconds")]
+
+# Only keep first entry
+dat1 <- subset(dat1, !duplicated(dat1$subj_id))
+dat3 <- subset(dat3, !duplicated(dat3$subj_id))
+
+# Remove empty variables wave 1
+which(apply(dat1, 2, function(x) is.na(x) |> sum()) == nrow(dat1)) |> names()
+dat1 <- subset(dat1, select = !grepl("^Q1.*", names(dat1)))
+
+# Check for other variables that are empty, but are supposed to be filled
+empty_var_w1 <- which(apply(dat1, 2, function(x) is.na(x) |> sum()) == nrow(dat1)) |>
+  names() |> print()
+empty_var_w2 <- which(apply(dat2, 2, function(x) is.na(x) |> sum()) == nrow(dat2)) |>
+  names() |> print()
+empty_var_w3 <- which(apply(dat3, 2, function(x) is.na(x) |> sum()) == nrow(dat3)) |>
+  names() |> print()
+empty_var_w4 <- which(apply(dat4, 2, function(x) is.na(x) |> sum()) == nrow(dat4)) |>
+  names() |> print()
+empty_var_w5 <- which(apply(dat5, 2, function(x) is.na(x) |> sum()) == nrow(dat5)) |>
+  names() |> print()
+empty_var_w6 <- which(apply(dat6, 2, function(x) is.na(x) |> sum()) == nrow(dat6)) |>
+  names() |> print()
+
+# Problems in wave 1
+# Theory: *After* data collection survey was altered and a new variable was
+# created. Data presentation was correct. (Angelica and Nora checked it via view
+# response in Qualtrics.)
+#
+# delg_tsk_typs_3 needs to be deleted
+# delg_tsk_typs_4 --> delg_tsk_typs_3
+# delg_tsk_typs_5 --> delg_tsk_typs_4
+# delg_tsk_typs_6 --> delg_tsk_typs_5
+# delg_tsk_typs_7 --> delg_tsk_typs_6
+# delg_tsk_typs_8 --> delg_tsk_typs_7
+
+# Rename variables as above
+dat1$delg_tsk_typs_3 <- dat1$delg_tsk_typs_4
+dat1$delg_tsk_typs_4 <- dat1$delg_tsk_typs_5
+dat1$delg_tsk_typs_5 <- dat1$delg_tsk_typs_6
+dat1$delg_tsk_typs_6 <- dat1$delg_tsk_typs_7
+dat1$delg_tsk_typs_7 <- dat1$delg_tsk_typs_8
+
+dat1$delg_tsk_typs_8 <- NULL
+
+# fav_Icecream is a bot detection item --> should it be removed?
+
+
+# Read codebook and rename variables for all waves
+cb <- openxlsx::read.xlsx("../HMC_codebook.xlsx")
+
+rename_data <- function(data, wave) {
+  data <- data[, cb$qualtricsname[cb[[wave]] == "x" | is.na(cb[[wave]])]]
+  names(data) <- cb$varname[cb[[wave]] == "x" | is.na(cb[[wave]])]
+  data
+}
+
+df1 <- rename_data(dat1, "wave1")
+df2 <- rename_data(dat2, "wave2")
+df3 <- rename_data(dat3, "wave3")
+df4 <- rename_data(dat4, "wave4")
+df5 <- rename_data(dat5, "wave5")
+df6 <- rename_data(dat6, "wave6")
+
+# identical(dat1$tsk_typ_if_wrtng_gnr_1, df1$tsk_typ_if_wrtng_1)
+
+
+save_dataframe <- function(data, wave, path = "../03_cleaned_data/") {
+  # as CSV
+  write.csv(data, 
+    paste0(path, "/HMC_", wave, "_cleaned.csv"),
+           row.names = FALSE)
+  # # as XLSX
+  # openxlsx::write.xlsx(data,
+  #   paste0(path, "/HMC_", wave, "_cleaned.xlsx"),
+  #   rowNames = FALSE)
+  # # as SAV
+  # haven::write_sav(data,
+  #   paste0(path, "/HMC_", wave, "_cleaned.sav"))
+}
+
+# Save anonymized data frames
+save_dataframe(df1, "wave1")
+save_dataframe(df2, "wave2")
+save_dataframe(df3, "wave3")
+save_dataframe(df4, "wave4")
+save_dataframe(df5, "wave5")
+save_dataframe(df6, "wave6")
+