Extract coordinates out of list() in R (bbox) and get centroid of bbox

Question

I have a data.frame with a column called geo. In this column, there are lists with coordinates that represent a bounding box. I would like to grab the coordinates that are inside each list and calculate the centroid for each bbox (new column with centroid as a point).

dput:

structure(list(full_name = c("Karlsruhe, Deutschland", "Volksparkstadion", 
"Hamburg, Deutschland"), country_code = c("DE", "DE", "DE"), 
    place_type = c("city", "poi", "city"), name = c("Karlsruhe", 
    "Volksparkstadion", "Hamburg"), country = c("Deutschland", 
    "Deutschland", "Deutschland"), id = c("5b146bf0b819f1af", 
    "0fc3a3fa5494c000", "5bcd72da50f0ee77"), geo = list(bbox = list(
        8.2773, 48.9405, 8.5418, 49.0914), bbox = list(9.8986, 
        53.5871, 9.8986, 53.5871), bbox = list(8.4202, 53.3951, 
        10.3252, 53.9647))), row.names = c(2L, 5L, 8L), class = "data.frame")

Answer 1

You can iterate the list contained in the data.frame using lapply. I am nesting a bunch of functions st_coordinates(st_centroid(st_as_sfc(st_bbox())), which you could plausibly pipe (eg., magrittr %>% or native R |>), to return the centroids then using do.call on the resulting list to create a data.frame that can be added back to your data.

library(sf)

dat <- structure(list(full_name = c("Karlsruhe, Deutschland", "Volksparkstadion", 
"Hamburg, Deutschland"), country_code = c("DE", "DE", "DE"), 
    place_type = c("city", "poi", "city"), name = c("Karlsruhe", 
    "Volksparkstadion", "Hamburg"), country = c("Deutschland", 
    "Deutschland", "Deutschland"), id = c("5b146bf0b819f1af", 
    "0fc3a3fa5494c000", "5bcd72da50f0ee77"), geo = list(bbox = list(
        8.2773, 48.9405, 8.5418, 49.0914), bbox = list(9.8986, 
        53.5871, 9.8986, 53.5871), bbox = list(8.4202, 53.3951, 
        10.3252, 53.9647))), row.names = c(2L, 5L, 8L), class = "data.frame")
        
( centroid <- lapply(dat$geo, function(x) {
  st_coordinates(st_centroid(st_as_sfc(st_bbox(c(xmin=x[[1]], 
      xmax=x[[2]], ymin=x[[3]], ymax=x[[4]])))))
}) )

( dat <- data.frame(dat, do.call('rbind', centroid)) )

Piped version (may be more understandable to tidyverse users) of the lapply step using the R native pipe so, not very tidy.

centroid <- lapply(dat$geo, function(x) {
  st_bbox(c(xmin=x[[1]], xmax=x[[2]], 
        ymin=x[[3]], ymax=x[[4]])) |> 
    st_as_sfc() |>
      st_centroid() |> 
        st_coordinates() 
})