---
title: "Lifebrain Global health survey"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{gbhs}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.retina = 3,
  out.width = "100%",
  fig.width = 8
)
ggplot2::theme_set(
  ggplot2::theme_classic() + 
    ggplot2::theme(legend.position = "bottom")
  )
```

Between June 2019 and August 2020, Lifebrain conducted the Global Brain Health Survey to collect data on people’s perceptions of brain health and willingness to take care of their brain by adopting new lifestyles. The survey was conducted online and translated into 14 languages to reach as many people as possible. In total, it collected 27,590 responses from people in 81 countries.

Please read more about the survey, its publications and the collaborators we worked with on [the Lifebrain webpates](https://www.lifebrain.uio.no/global-brain-health-survey/).

This package contains the cleaned data from this survey, along with tools to access the raw data, and scripts used to create reports for publication. 

## Accessing and working with the data

The data can be loaded as most R users would be used to loading built-in package data.
When the package is installed, you can call the `gbhs` object, and view the data.

```{r setup}
library(tidyverse)
library(gbhs)
gbhs
```

The data has mainly been prepared with the [tidyverse]() collection of packages, and as such, we will continue showcasing a couple of things that you could do with the data.

```{r}
library(dplyr)

# Have only answers from Norway
gbhs |> 
  filter(q19_residence == "Norway")

# Have only answers from Norway
# And key variables + question 2
gbhs |> 
  filter(q19_residence == "Norway") |> 
  select(1:3, starts_with("q02"))
```
## Turning groups of questions to long format
There are several clusters of questions probing various concepts around brain health. 
These are linked together through naming of columns.
For easy of use in plotting and analysis, we created a convenience function that will turn the data longer based on these clusters.
When turning the data long with `gbhs_long_q()` you need to provide the GBHS data to the function, along with indication on which cluster of questions to take.
In the example below, we turn the answers to question 2 clusters longer.
This will add another set of columns to the data:
- `key` : what was previously column names
- `value`: what was previously cell values
- `binary`: binary classification of `value` if value is not already binary.

```{r fig.height=12}
library(ggplot2)
gbhs |> 
  gbhs_long_q(2) |> 
  ggplot(aes(y = value, group = q18_age, fill = q18_age)) +
  geom_bar() +
  theme_minimal() +
  facet_wrap(~key)
```


```{r fig.height=12}
gbhs |> 
  gbhs_long_q(11) |> 
  ggplot(aes(y = value)) +
  geom_bar() +
  theme_minimal() +
  facet_wrap(~key)
```


## Using the custom plotting functions

If you want to reproduce the types of plots we create, we have also provided these in the package.

```{r}
gbhs |> 
  gbhs_long_q(11) |> 
  ggbar(age) +
  theme(axis.text.x = element_text(angle = 30, hjust = 1))
```

```{r}
gbhs |> 
  gbhs_long_q(2) |> 
  ggstacked( 
    pattern = "No|Mod|We",
    n_breaks = 5,
  ) +
  theme(legend.position = "right") +
  coord_cartesian(xlim = c(0, 1.1), clip = "off") 
```

```{r, fig.height=12}
gbhs |> 
  gbhs_long_q(2) |> 
  ggstacked_wrap(
    y = age,
    pattern = "No|Mod|We",
    n_breaks = 5,
  ) +
  theme(legend.position = "right") +
  coord_cartesian(xlim = c(0, 1.1), clip = "off") 
```


## Inspecting meta-data
There are several pieces of meta-data available. 
The full meta-data provides the entire json that the nettskjema API (which was used to retrieve survey data) can provide. 
It is a full list of all questions, formulations, order and coding as the nettskjema survey platform used to run the survey.
The json in its entirety is included in the package files, and its location on your system can be found by

```{r}
gbhs_path_meta(type = "meta-data")
gbhs_path_meta("119161_sp.json", type = "meta-data")

gbhs_path_meta("119161_sp.json", type = "meta-data") |> 
  readLines(n = 15) |> 
  cat(sep = "\n")
```

While the codebooks are contained in the meta-data, these can also be retrieved in two ways.
Either as json, or as a tidied csv.

```{r}
gbhs_path_meta(type = "codebook")
gbhs_path_meta("119161_sp.json", type = "codebook")

gbhs_path_meta("119161_sp.json", type = "codebook") |> 
  readLines(n = 15) |> 
  cat(sep = "\n")

gbhs_path_meta("119161_sp.tsv", type = "codebook") |> 
  read.delim() |> 
  knitr::kable()
```