idemsinternational / parentappdatascripts

Some R-Scripts for the cleaning and use of ParentApp data

ParentAppDataScripts

Some R-Scripts for the cleaning and use of ParentApp data

parentappdatascripts's Issues

Using `select` to group categories in the data set for tables

Currently in the Metabase analysis code, we create subsets of data to then run tables.
We want to use the select function instead.

For example,

data_baseline_survey <- with(plhdata_org_clean, data.frame(app_user_id,
                                                           Org,
                                                           rp.contact.field.survey_welcome_completed, rp.contact.field.user_gender,
                                                           rp.contact.field.user_age, rp.contact.field.household_adults, 
                                                           rp.contact.field.household_teens, rp.contact.field.household_babies,
                                                           rp.contact.field.household_children))

colnames(data_baseline_survey) <- naming_conventions(colnames(data_baseline_survey), replace = "rp.contact.field.")
two_way_table_baseline <- NULL
# to run all the "Habits" tables individually:
for (i in 3:length(data_baseline_survey)){
  two_way_table_baseline[[(i-2)]] <- two_way_table(data = data_baseline_survey,
                                                   row_var = names(data_baseline_survey)[i])
  
  # to get user IDs
  # user_id_print(names(data_baseline_survey)[i])
}
names(two_way_table_baseline) <- names(data_baseline_survey)[3:length(data_baseline_survey)]
# then to access a table:
two_way_table_baseline$`Household babies`

Can become

baseline_survey_columns <- c("rp.contact.field.survey_welcome_completed", "rp.contact.field.user_gender",
                             "rp.contact.field.user_age", "rp.contact.field.household_adults", 
                             "rp.contact.field.household_teens", "rp.contact.field.household_babies",
                             "rp.contact.field.household_children")

two_way_table_baseline <- NULL
for (i in 1:length(baseline_survey_columns)){
  two_way_table_baseline[[i]] <- two_way_table(data = plhdata_org_clean,
                                               row_var = baseline_survey_columns[i],
                                               replace = "rp.contact.field.")
}
names(two_way_table_baseline) <- baseline_survey_columns

Presumably we also want to use a map function instead of the for loop. For example,

two_way_table_baseline <- plhdata_org_clean %>%
  dplyr::select(Org, baseline_survey_columns) %>%
  purrr::map_at(baseline_survey_columns,
                two_way_table(row_var = names(.), replace = "rp.contact.field."))

(with the minor exception that I have used purrr incorrectly as it is giving an error - need to work this one out)

Renaming variables and adding variable labels to the data set

To tidy up the data set, it would be good to rename the variables. For example, for

"rp.contact.field.survey_welcome_completed"

to become

"Survey welcome completed"

This can currently be achieved with the naming_conventions function. However, does that lose information in the variable? Would we therefore want to add variable labels to the data?

Adding percentages to two-way tables

Could write in percentages for two-way tables function (as with multiple_summary_PT function) with the option of giving them by row or by column.

Row would be useful for e.g. gender (what percentage of Nontobeko users were male vs female)
Column could give info on share by organisation (e.g. what percentage of all users who have completed a workshop are from organisation x), but is probably less relevant.