Spring 2026 Slides

class: inverse, bottom
background-image: url(https://images.unsplash.com/photo-1588591795084-1770cb3be374?q=80&w=2070&auto=format&fit=crop&ixlib=rb-4.1.0&ixid=M3wxMjA3fDB8MHxwaG90by1wYWdlfHx8fGVufDB8fHx8fA%3D%3D)
background-size: cover

## .large[R Medicine 2026: ggregions for building intuitive coding interfaces for anatomical visualizations]
## .small[]
#### .tiny[Dr. Evangeline Reynolds | 2026-04-04 | DU, Image credit: Ross Sneddon, Upsplash]

???

Title slide

---
class: inverse

# Motivations for ggregions

> ### I am told there are people who do not care for maps, and I find it hard to believe. - Robert Louis Stevenson

> ### I hate when I work with spatial data that so much of my mindshare goes to thinking about 'how do I wrangle this this map and what on earth is going on' versus the actual meaning behind what I'm trying to map. - Emily Riederer

> ### 'My teacher, Mr. Jayson, says a map is a picture of someplace from above.  It's like flying over that spot in an airplane.' - 'Lisa' in Loreen Leedy's in 'Mappying Penny's World'

---
class: inverse

## Maps are some of the most compelling and intuitive data visualizations, 
--
and they are some of the first visualizations people will come across in their lives.

## And there is great support for mapping in R and in the data visualization library {ggplot2} which supports many R packages.

## But, producing a simple map, like a choropleth, can feel much harder than producing other plots, like a scatterplot for example in ggplot2.

---

``` r
*ggplot(cars)   # data
```

---
count: false
## A scatter plot is produced by:

``` r
ggplot(cars) +  # data
*  aes(x = speed, y = dist)   # position and other visual channel mapping
```

---
count: false
## A scatter plot is produced by:

``` r
ggplot(cars) +  # data
   aes(x = speed, y = dist) +  # position and other visual channel mapping
*  geom_point()  # layer
```

???

## In other words, you might say, 'used this data, and points, and the positions are defined by x and y'

---

## ggregions proposed API

``` r
ggplot(mapping_data) +             # data
*  aes(region_id = var1_region_id,  # position mapping
       fill = var2) +               # and other visual channel mapping
*  geom_region()                    # layer
```

> ## [ggregions-generated] APIs feels like they actually puts the "question" at the center. - Emily Riederer

---

# What {ggregions} delivers

## Effortlessly write new region-specific geom_* functions.

---

## Example with geographic regions: US states

### Step 1. prepare data

``` r
library(dplyr)

# Step 1. prep some of their geo reference data
us_states_ref <- usmapdata::us_map() |>
  select( state_name = full, # state_name will positional aes like x or y 
          state_abbr = abbr, 
          fips,
          geometry = geom,  # one column with geometry is required
         )  # state_abbr will also be optional positional aes like x or y
```

---

### Step 2. write user-facing functions with ggregion 'write' utilities

``` r
# Step 2. use `write_geom_region_locale`
library(ggregions)
geom_us_state <- write_geom_region_locale(us_states_ref)
stamp_us_state <- write_stamp_region_locale(us_states_ref)
geom_us_state_text <- write_geom_region_text_locale(us_states_ref)
stamp_us_state_text <- write_stamp_region_text_locale(us_states_ref)
```

---

### Step 2b. (optional) Inspect user-facing function

``` r
geom_us_state
```

```
function (mapping = aes(), data = NULL, stat = StatRegion, position = "identity", 
    na.rm = FALSE, show.legend = NA, inherit.aes = TRUE, ref_data = us_states_ref, 
    ...) 
{
    c(layer_sf(geom = GeomSf, data = data, mapping = mapping, 
        stat = stat, position = position, show.legend = show.legend, 
        inherit.aes = inherit.aes, params = rlang::list2(na.rm = na.rm, 
            ref_data = ref_data, ...)), coord_sf(crs = st_crs_mod(ref_data)))
}
<environment: namespace:ggregions>
```

---

#### Steb 2bb (for the truely curious) Inspect StatRegions compute

``` r
ggregions:::StatRegion$compute_panel
```

```
<ggproto method>
  <Wrapper function>
    function (...) 
compute_panel(...)

<Inner function (f)>
    function (data, scales, ref_data, keep = NULL, drop = NULL, stamp = F) 
{
    ref_data$id <- ref_data[1][[1]]
    if (!is.null(keep)) {
        ref_data <- dplyr::filter(ref_data, id %in% keep)
    }
    if (!is.null(drop)) {
        ref_data <- dplyr::filter(ref_data, !(id %in% drop))
    }
    ref_data <- ggplot2::StatSfCoordinates$compute_group(ggplot2::StatSf$compute_panel(ref_data, 
        coord = ggplot2::CoordSf), coord = ggplot2::CoordSf)
    if (!stamp) {
        dplyr::inner_join(ref_data, data)
    }
    else {
        ref_data
    }
}
```

---

``` r
*library(ggplot2)
```
]
 
.panel2-usmapdata-auto[

]

---
count: false

``` r
library(ggplot2)
*library(usmapdata)
```
]
 
.panel2-usmapdata-auto[

]

---
count: false