---
title: "Data Sources"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Data Sources}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

We provide access to a variety of different data sources in `conmat`. Most of these are centred around Australian data, as the package was initially created for disease modelling work in Australia. The aim of this vignette is to give a quick tour of the data sources available in `conmat`.

```{r setup}
library(conmat)
```

## World data

We provide functions to clean up world population data from `socialmixr`.

```{r}
world_data <- socialmixr::wpp_age()

head(world_data)
```

We can tidy the data up, filtering down to a specified location and year with the `age_population` function:

```{r}
nz_2015 <- age_population(
  data = world_data,
  location_col = country,
  location = "New Zealand",
  age_col = lower.age.limit,
  year_col = year,
  year = 2015
)

nz_2015
```

This returns a `conmat_population` object, which is a data frame that knows which columns represent `age` and `population` information. This is useful for other modelling parts of the `conmat` package.

## Australian Bureau of Statistics (ABS) data

### Accessing Functions

We provide two functions to access LGA  (Local Government Area), and state level population age data, which are provided in 5 year age bins from 0, 5, up to 85+. These data are `conmat_population` tibbles, which means that they know which columns represent the `age` and `population` information. This means that functions inside of `conmat` can work a bit smoother as we refer to these columns frequently.

#### `abs_age_lga()`

```{r abs-age-lga}
fairfield <- abs_age_lga(lga_name = "Fairfield (C)")
fairfield
```

Note that this is a `conmat_population` object, which prints in red at the top of the data frame. This provides the information on the `age` and `population` columns, stating: `age: lower.age.limit`, and `population: population`, indicating which columns refer to the appropriate variables.

Also note that `abs_age_lga` requires you to know the exact name of the LGA, you can see them in the dataset, `abs_lga_lookup`

```{r abs-lga-lookup}
abs_lga_lookup
```

And if you're not sure about a particular name of a place, you can use `agrep` and `filter`, to match on similar-ish characters, like so:

```{r abs-lga-lookup-filter}
library(dplyr)
abs_lga_lookup %>%
  filter(agrepl("Sydney", lga))
```

#### `abs_age_state()`

This takes in the abbreviated state names, and is also a `conmat_population` object.

```{r abs-age-state}
abs_age_state(state_name = "NSW")
```

You can see these state names with:

```{r}
unique(abs_lga_lookup$state)
```

Note that "OT" stands for "other territories"

### ABS data

We provide other ABS data, listed now. You can read the full details of the data at their respective helpfiles, by writing, for example, `?abs_education_state`.

#### Education by state data for 2006 -2020

```{r abs-education-state}
abs_education_state
```

#### Education by state data for 2020

```{r abs-education-state-2020}
abs_education_state_2020
```


#### Employment by LGA for 2016

```{r abs-employ-age-lga}
abs_employ_age_lga
```


#### Number of people in each household by LGA for 2016

```{r abs-household-lga}
abs_household_lga
```


#### LGA age population for 2016 for all states and LGAs

```{r abs-pop-age-lga-2016}
abs_pop_age_lga_2016
```


#### LGA age population for 2020 for all states and LGAs

```{r abs-pop-age-lga-2020}
abs_pop_age_lga_2020
```

#### State age population for 2020

```{r abs-state-age}
abs_state_age
```


## Epidemiology / disease modelling data

### Transmission probabilities from Eyre

A dataset containing data digitised from "The impact of SARS-CoV-2 vaccination on Alpha & Delta variant transmission", by David W Eyre, Donald Taylor, Mark Purver, David Chapman, Tom Fowler, Koen B Pouwels, A Sarah Walker, Tim EA Peto

```{r eyre-transmission-probabilities}
eyre_transmission_probabilities
```

We can visualise the data like so:

```{r eyre-transmission-probabilities-plot}
library(ggplot2)
library(stringr)
library(dplyr)
eyre_transmission_probabilities %>%
  group_by(
    setting,
    case_age_5y,
    contact_age_5y
  ) %>%
  summarise(
    across(
      probability,
      mean
    ),
    .groups = "drop"
  ) %>%
  rename(
    case_age = case_age_5y,
    contact_age = contact_age_5y
  ) %>%
  mutate(
    across(
      ends_with("age"),
      ~ factor(.x,
        levels = str_sort(
          unique(.x),
          numeric = TRUE
        )
      )
    )
  ) %>%
  ggplot(
    aes(
      x = case_age,
      y = contact_age,
      fill = probability
    )
  ) +
  facet_wrap(~setting) +
  geom_tile() +
  scale_fill_viridis_c() +
  coord_fixed() +
  theme_minimal() +
  theme(
    axis.text = element_text(angle = 45, hjust = 1)
  )
```