survey-data.Rmd

---
title: "PLAY Survey Data"
author: "Rick Gilmore"
date: "`r Sys.time()`"
output: 
    html_document:
      self_contained: yes
      code_folding: show
      toc: true
      toc_depth: 2
      toc_float: true

params:
  db_login: email@provider.com
  play_volume_id: 1280
  play_survey_session_id: 51539
---

# Purpose

This document shows how to access and visualize survey response data from the PLAY dataset.

# Rendering

To render the document, issue this command from the R console

`rmarkdown::render('survey-data.Rmd', params = list(db_login='<YOUR_DATABRARY_LOGIN>'))`

substituting your actual Databrary login (email account) for `<YOUR_DATABRARY_LOGIN>`.

If you try to 'knit' the document another way, it won't work.

# Setup

There are several R packages we need to run this code.
The following will test to see if the packages are installed on your system.
If not, the packages will be installed for you.

```{r setup, include=TRUE}
knitr::opts_chunk$set(echo = TRUE)

if (!require(tidyverse)) {
  install.packages('tidyverse')
}

if (!require(devtools)) {
  install.packages('devtools')
}

if (!require(databraryapi)) {
  devtools::install_github('PLAY-behaviorome/databraryapi')
}

library(tidyverse) # for pipe `%>%` operator
```

You will need to login to Databrary in order to access the PLAY data.
The next chunk asks for your Databrary credentials.
It returns `TRUE` if your login is successful.

```{r login-databrary}
databraryapi::login_db(params$db_login)
```

# List files

Databrary 1.0 has a volume/session structure.
A volume is a dataset or collection.
A session is a collection of related files.
You can think of a session as a folder.
The survey data are in a session folder called 'KoBoToolbox Survey Data'.
Volumes and sessions have ID numbers that Databrary uses to access each.

The PLAY data release volume has an id of `r params$play_volume_id`.
The survey data session has a session id of `r params$play_survey_session_id`.

So, to go to the PLAY volume in your browser, you can visit <https://nyu.databrary.org/volume/1280>.

The following code, however, does all of this within an R session.

```{r load-play-survey-files}
play_survey_files <-
  databraryapi::list_assets_in_session(session_id = params$play_survey_session_id)

# Print the files data in a 'pretty' way
play_survey_files %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(., bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

Note that each file has an `asset_id`.
We'll use this to access the data file(s) we want.

# Import by measure

## Pets

```{r load-pets}
# Grab the asset_id for the pets data based on the name
pets_in_name <- stringr::str_detect(play_survey_files$name, 'pets')
pets_asset_id <- play_survey_files$asset_id[pets_in_name]

# Select and download the pets data
pets <-
  databraryapi::read_csv_data_as_df(params$play_survey_session_id, as.numeric(pets_asset_id))

# Print the pets data
pets %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(., bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

We can do a bit of clean-up here to see what families have pets and of what type.

```{r clean-pets}
pets_cleaned <- pets %>%
  dplyr::mutate(., pets_at_home = recode(pets_at_home, 'Sí' = 'Yes'),
                has_dog = stringr::str_detect(type_number, '[dD]og'),
                has_cat = stringr::str_detect(type_number, '[cC]at'),
                where_live = recode(where_live, 'Adentro' = 'Indoors')) %>%
  dplyr::arrange(., child_age_grp)
```

### Homes with dogs

```{r homes-w-dogs}
pets_cleaned %>%
  dplyr::filter(., has_dog == TRUE) %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(., bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

It looks like there are more homes with dogs.

### Homes with cats

```{r homes-w-cats}
pets_cleaned %>%
  dplyr::filter(., has_cat == TRUE) %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(., bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

Most cats and dogs live indoors.
This must be in NYC!

Also, there seem to be dog families and cat families, but not both.

## Typical day

```{r typical-day}
# Grab the asset_id for the data based on the name
typical_in_name <-
  stringr::str_detect(play_survey_files$name, 'typical')
typical_asset_id <- play_survey_files$asset_id[typical_in_name]

# Select and download the data
typical_day_data <-
  databraryapi::read_csv_data_as_df(params$play_survey_session_id, as.numeric(typical_asset_id))

# Print the data
typical_day_data %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

We'll do a little clean-up and select a smaller number of columns.

```{r clean-typical-day}
typical_cleaned <- typical_day_data %>%
  dplyr::mutate(
    .,
    typical_day = recode(typical_day, 'Sí' = 'Yes'),
    activities_similar = recode(activities_similar, 'Sí' = 'Yes'),
    typical_night_morning = recode(typical_night_morning, 'Sí' = 'Yes'),
    unusual_feelings = recode(unusual_feelings, 'Sí' = 'Yes')
  ) %>%
  dplyr::select(
    .,
    play_site_id,
    child_age_grp,
    typical_day,
    activities_similar,
    typical_night_morning,
    unusual_feelings
  ) %>%
  dplyr::arrange(., child_age_grp)

# Print the data
typical_cleaned %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

Note that the `unusual_feelings` variable might need to be examined more closely.
This short name might not reflect the actual question properly.

```{r}
typical_day %>%
  dplyr::select(., child_age_grp, typical_day) %>%
  dplyr::filter(., !is.na(typical_day)) %>%
  dplyr::mutate(., typical_day = recode(typical_day, 'Sí' = 'Yes')) %>%
  dplyr::group_by(child_age_grp) %>%
  dplyr::summarise(n_families = n(),
                   n_rating_typical = sum(typical_day == 'Yes')) %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "250px")
```

## Household labor

```{r household-labor}
# Grab the asset_id for the data based on the name
labor_in_name <-
  stringr::str_detect(play_survey_files$name, 'labor')
household_labor_asset_id <- play_survey_files$asset_id[labor_in_name]

# Select and download the data
household_labor <-
  databraryapi::read_csv_data_as_df(params$play_survey_session_id, as.numeric(household_labor_asset_id))

# Print the data
household_labor %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

## Locomotor milestones

```{r loco-milestones}
# Grab the asset_id for the data based on the name
loco_in_name <-
  stringr::str_detect(play_survey_files$name, 'loco')
loco_milestones_asset_id <- play_survey_files$asset_id[loco_in_name]

# Select and download the data
loco_milestones <-
  databraryapi::read_csv_data_as_df(params$play_survey_session_id, as.numeric(loco_milestones_asset_id))

# Print the data
loco_milestones %>%
  # Drop comments here for privacy...
  dplyr::select(., -c('WHO_walk_comments', 'walk_onset_comments', 'crawl_onset_comments')) %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

Clean-up to focus on milestone ages.

```{r}
loco_cleaned <- loco_milestones %>%
  dplyr::select(.,
                child_age_grp,
                child_sex,
                WHO_walk_month,
                walk_onset_month,
                crawl_onset_month) %>%
  tidyr::pivot_longer(
    .,
    cols = c('WHO_walk_month', 'walk_onset_month', 'crawl_onset_month'),
    names_to = "milestone",
    values_to = "age_mos"
  )

loco_cleaned %>%
  ggplot(.) +
  aes(x = age_mos, fill = child_sex) +
  geom_histogram() +
  facet_grid(milestone ~ .)
```

## Health

```{r health}
# # Grab the asset_id for the data based on the name
# health_in_name <-
#   stringr::str_detect(play_survey_files$name, 'health')
# health_asset_id <- play_survey_files$asset_id[health_in_name]
# 
# # Select and download the data
# health <-
#   databraryapi::read_csv_data_as_df(params$play_survey_session_id, as.numeric(health_asset_id))
# 
# # Print the data
# health %>%
#   kableExtra::kbl(.) %>%
#   kableExtra::kable_styling(.,
#                             bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
#   kableExtra::scroll_box(width = "800px", height = "300px")
```


## Media

```{r media}
# Grab the asset_id for the data based on the name
media_in_name <-
  stringr::str_detect(play_survey_files$name, 'media')
media_asset_id <- play_survey_files$asset_id[media_in_name]

# Select and download the data
media <-
  databraryapi::read_csv_data_as_df(params$play_survey_session_id, as.numeric(media_asset_id))

# Print the data
media %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "300px")
```

Do some clean-up to summarize some of the findings.

```{r media-summarize}
media %>%
  dplyr::select(., child_age_grp, have_tv, child_used_tv) %>%
  dplyr::filter(., !is.na(child_used_tv)) %>%
  dplyr::mutate(., child_used_tv = recode(child_used_tv, 'Sí' = 'Yes')) %>%
  dplyr::group_by(child_age_grp) %>%
  dplyr::summarise(homes_w_tv = sum(have_tv),
                   children_use_tv = sum(child_used_tv == "Yes")) %>%
  kableExtra::kbl(.) %>%
  kableExtra::kable_styling(.,
                            bootstrap_options = c("striped", "hover", "condensed", "responsive")) %>%
  kableExtra::scroll_box(width = "800px", height = "250px")

```




# Clean-up

Log out of Databrary.

```{r logout-databrary}
databraryapi::logout_db()
```