json_to_tibble.R

library(renv)
renv::restore()
library(jsonlite)
library(tibble)
library(dplyr)
library(tidyr)
library(purrr)
library(stringr)
raw <- fromJSON("./data_export.20250228.json")
df <- tibble(raw) %>%
  mutate(
    patient = names(raw)
  )
df <- unnest(df, raw) %>%
  mutate(dimension = names(raw)) %>%
  pivot_wider(names_from = dimension, values_from = raw)
# Discard non-CRF data
df <- select(df, c(patient, external_crf_data))
df <- mutate(
  df,
  pecunia = map(external_crf_data, function(d) {
    if (!is_null(d)) {
      filter(d, crf_id == "PECUNIA")
    }
  })
)
df <- select(df, -external_crf_data)
df <- mutate(
  df, 
  pecunia = map(pecunia, function(d) {
    if (!is_null(d)) mutate(d, json_data = tibble(json_data))
  })
) %>% unnest(pecunia)
df <- unnest(df, json_data)
df <- mutate(
  df, 
  items = map(items, function(d) {
    as_tibble(d)
  })
) %>% unnest(items)

# Data
head(df)


# stats
d <- filter(df, !str_starts(patient, 'TST'))

# Number of responses
d %>% select(
  c(patient, crf_incremental_id)
) %>% unique() %>%
  group_by(
    patient
  ) %>%
  summarise(
  times_completed = n()
) %>%
  group_by(times_completed) %>%
  summarise(
    patients = n()
  )