====== REDCap Data ======

We use the API to download redcap data to csv and RData files that can easily be ingested by other scripts.

This is done automatically by [[:tools:cron]] on [[:admin:it:rhea]] (as ''foranw@rhea'')

  * ''/Volumes/Hera/Projects/Habit/behave/redcap/00_get.R'' -> ''data/habit_redcap.csv''
  * '':/Volumes/Hera/Projects/SPA/scripts/redcap/00_get.R '' -> ''spa_redcap.RData''

===== Data Shape =====

As exported, data is in "fake wide" format. Rows are repeated by event (`behave`,`mr`,`behave_2`, `mr_2`, etc) but columns exist for all field labels for all events.

For instance, ''recuit_age'' is only presented during ''recruit'', so it is ''NA'' for all other ''redcap_event_name''s

^ record_id ^ redcap_event_name ^ recruit_age ^ enroll_date ^ demo_age ^ ptx_mp2rage_onset   ^
|     11751 | recruit_arm_1     |          23 |             |          |                     |
|     11751 | screener_arm_1    |             |             |          |                     |
|     11751 | behavioral_arm_1  |             |  2023-12-06 |          |                     |
|     11751 | mr_arm_1          |             |             |          | 2023-12-08 09:06:00 |


To make truly wide, subset columns for each event name of interest and merge back. Example pulled from ''/Volumes/Hera/Projects/SPA/scripts/redcap/age_visit_histogram.R''

<code=r>
# remove '_arm_1' parse visit number from event name
rm_arm <- function(d)
  d |> mutate(redcap_event_name=gsub('_arm_1','',redcap_event_name),
              visitno=str_extract(redcap_event_name, '[0-9]$'),
              visitno=ifelse(is.na(visitno), 1, visitno))

d.behave <- lunas |>
  select(luna=record_id, redcap_event_name,
         age=demo_age, behave=behave_starttime) |>
  filter(!is.na(behave)) |>
  rm_arm()
d.mri <- lunas |>
  select(luna=record_id, redcap_event_name,
         ptx=ptx_mp2rage_onset)|>
  filter(!is.na(ptx)) |>
  rm_arm()

d_wide <- left_join(d.behave, d.mri, by=c("luna","visitno"))

d_wide |>
  rename_with(\(x) gsub('redcap_event_name','name', x)) |>
  filter(luna %in% c(11751,11823))
</code>

^  luna ^ name.x     ^ age ^ behave              ^ visitno ^ name.y ^ ptx                 ^
| 11751 | behavioral |     | 2023-12-06 08:21:39 |       1 | mr     | 2023-12-08 09:06:00 |
| 11823 | behavioral |  13 | 2024-08-09 14:51:18 |       1 |        |                     |