eta.qmd

```{r setup, include = FALSE}
source("global.R")
```


# ETA plots

## Setup

We start by defining a set of ETAs to use in the plots. 

```{r}
etas <- c(
  "ETA1//ETA-CL", 
  "ETA2//ETA-V2", 
  "ETA3//ETA-KA"
)
```

This is in the `col-label` format described earlier. We also set out a set 
of covariates that we can use for ETA diagnostics.

```{r}
covs <- c(
  "WT//Weight (kg)", 
  "ALB//Albumin (g/dL)", 
  "SCR//Creatinine (mg/dL)"
)
```

## Data used on this page

We are exclusively using a data set that is one row per individual

:::{.callout-note collapse=true appearance="simple"}

## Preview the data used on this page

```{r}
head(as.data.frame(id), n=3)
```

:::


## Versus continuous `[eta_cont]`

Grouped by eta

```{r}
#| fig-width: 6
#| fig-height: 6
eta_cont(id, x = covs, y = etas[2]) %>% 
  pm_grid()
```

Grouped by covariate

```{r}
#| fig-width: 6
#| fig-height: 6
eta_cont(id, x = covs[1], y = etas) %>% 
  pm_grid(ncol = 2)
```

## By categorical `[eta_cat]`

```{r}
#| fig-width: 6
#| fig-height: 6
eta_cat(id, x = "STUDYc//Study type", y = etas) %>% 
  pm_grid()
```

## Histograms `[eta_hist]`

```{r}
#| fig-width: 6
#| fig-height: 6
eta_hist(id, etas, bins = 10) %>% 
  pm_grid()
```

## Pairs  `[eta_pairs]`

See also [Chapter -@sec-pairs] on making pairs plots. 

```{r}
#| fig-width: 6
#| fig-height: 5
eta_pairs(id, etas) 
```