R googleway with ABS census data

Question

I'm intending to analyse Australian census data using googleway to produce heat maps.

My approach has been to use prepared data from googleway melbourne which contains column SA2_NAME and join it with the ESRI shape file from the census data after conversion with rgdal (code below). The problem is that joining by SA2_NAME is not unique for polylines - some SA2 areas are made of multiple 'sub' areas. So it seems this is not a good approach.

A better approach would be to convert the ESRI shape data sa2_shape below to have polylines in the format of the melbourne data. How is this done?

Code below produces a 'bridging' data frame to use in joining melbourne data from googleway with ABS data which has SA2_MAIN as the key field - as stated above, the problem with this 'hack' approach is that polylines are not unique by SA2_NAME

library(tidyverse)
library(googleway)
library(rgdal)

shape_path <- "abs_data/sa2_esri_shapefile"
shape_file <- "SA2_2016_AUST"
sa2_shape <- readOGR(shape_path, shape_file)
sa2_df <- data.frame(sa2_shape$SA2_MAIN, sa2_shape$SA2_NAME)
names(sa2_df) <- c("SA2_MAIN", "SA2_NAME")
sa2_df <- sa2_df %>% semi_join(melbourne, by = "SA2_NAME")

If you use `sf` objects they will naturally get plotted by `googleway`. Use `library(sf); sf <- sf::st_read( 'path_to_shape_file.shp' )` to load your data as `sf` objects. — SymbolixAU, Aug 29 '18 at 08:09
Or, if you're using `sp` objects, convert them to `sf` using `sf::st_as_sf()` — SymbolixAU, Aug 29 '18 at 08:11
@SymbolixAU This works fine - also please note it seems google_map() crashes if the geometry column (from using sf) is an empty list. As a work around, this can be fixed by filtering these out. — Vlad, Aug 29 '18 at 12:54
You may be interested in the ASGS package [Note: 700MB] (for mapping SA2s): `library(ASGS.foyer); install_ASGS()` and the `Census2016.DataPack` package: https://github.com/hughparsonage/Census2016.Datapack — Hugh, Aug 29 '18 at 13:35
" crashes if the geometry column (from using sf) is an empty list " - are you able to give an example and add it to [the issue tracker](https://github.com/SymbolixAU/googleway/issues) ? I can't seem to generate this error. — SymbolixAU, Aug 29 '18 at 21:54
OK - added https://github.com/SymbolixAU/googleway/issues/176 — Vlad, Aug 30 '18 at 00:07

score 1 · Answer 1 · answered Aug 29 '18 at 13:08

As per SymbolixAU comment - used sf to load the data and this works as long as geometry is not an empty list - see code below.

library(tidyverse)
library(googleway)
library(sf)

shape_path <- "abs_data/sa2_esri_shapefile"
shape_file <- "SA2_2016_AUST"

shape_file_path <- paste0(shape_path, "/", shape_file, '.shp')
sa2_shape <- sf::st_read(shape_file_path)
sa2_shape <- sa2_shape %>% 
  filter(STATE_NAME == "Victoria",
         AREA_SQKM > 0)# This is important - otherwise google_map() will crash!
google_map() %>% 
  googleway::add_polygons(data = sa2_shape, 
                          polyline = "geometry", 
                          fill_colour = "SA2_NAME")

> sa2_shape %>% head()
Simple feature collection with 6 features and 6 fields
geometry type:  MULTIPOLYGON
dimension:      XY
bbox:           xmin: 143.6849 ymin: -37.68153 xmax: 143.951 ymax: -37.46847
epsg (SRID):    4283
proj4string:    +proj=longlat +ellps=GRS80 +towgs84=0,0,0,0,0,0,0 +no_defs
   SA2_MAIN SA2_MAIN16         SA2_NAME STATE_CODE STATE_NAME AREA_SQKM                       geometry
1 201011001  201011001        Alfredton          2   Victoria   52.7111 MULTIPOLYGON (((143.7072 -3...
2 201011002  201011002         Ballarat          2   Victoria   12.3787 MULTIPOLYGON (((143.8675 -3...
3 201011003  201011003 Ballarat - North          2   Victoria   92.3577 MULTIPOLYGON (((143.853 -37...
4 201011004  201011004 Ballarat - South          2   Victoria   32.8541 MULTIPOLYGON (((143.8675 -3...
5 201011005  201011005        Buninyong          2   Victoria   51.5855 MULTIPOLYGON (((143.8533 -3...
6 201011006  201011006        Delacombe          2   Victoria   34.1608 MULTIPOLYGON (((143.7072 -3...

R googleway with ABS census data

1 Answers1