Changelog • hpa.spatial

hpa.spatial (development version)

update my email in DESCRIPTION file.
fix problem caused by sf::st_crs() within update_crs() by importing {sf}.
create new vignette with example of more complex mapping of data using more up-to-date ERP dataset to adjust correspondence table creation.
add dataset (sa2_erp23) with estimated residential population (2023) at SA2 level.

remove hideous meta-programming that was handling the recursive calling of map_data_with_correspondence() (now that it’s not used recursively).
where possible to use a combination of ASGS correspondence tables (including combining multiple) use these for mapping data across editions/geographies.
avoid recursive programming for map_data_with_correspondence() by using new function: .map_data_with_ct()
update testing suite to be faster to run and more comprehensive.

data read from {hpa.spatial.data} is read from a specific commit to ensure (from now on) old versions of hpa.spatial will read the same data as they are expected to (did at the time of analyses).
allow user to specify a custom correspondence table to map_data_with_correspondence().
add access to recently added Aged Care Planning Region shape files (on {hpa.spatial.data}).
only do GHA for R CMD CHECK on Windows.
update README.
update vignettes.

Ensure that SA2s can completely map to LHNs (including SA2s with zero pop). Includes an associated change to {hpa.spatial.data} commit.

update create_child_geo() to assign un-parented mesh blocks to the nearest polygon from parent_geo.
update testing snaps after LHN polygon was updated on {hpa.spatial.data}.
create vignettes.

use open-source (MIT) license.
bug fix: when creating a correspondence table and there is a complete match between the polygons but all meshblocks data within have populations of zero, allow the match with ratio = 1.
bug fix: when using map_data_with_correspondence, avoid bug which converts Date grouping variables to numeric (use of cbind.data.frame, not cbind).
workflow: add github action workflows for checking package (check-standard) and making a github pages website (pkgdown).
bug fix: map_data_with_correspondence() would get confused when a vector object is used in combination with .data. Now the vector is attempted to be evaluated and if that errors, it accesses it from .data.

allow user to get Modified Monash Model data (for SA1, 2016 ed.) using get_polygon(name = "mmm19") or get_mmm19_poly().
bug fix: pass value_type appropriately when groups are used in map_data_with_correspondence().
feature: let map_data_with_correspondence() take a list of vectors for groups or, when .data is passed, can take a character vector or column names or a vector of symbols.

allow map_data_with_correspondence() to take from_geo and to_geo rather than only areas and years. (Also requires adding export_fname arg, consistent with changes to get_correspondence_tbl().)
allow get_correspondence_tbl() to take from_geo and to_geo rather than only areas and years to get the polygons via get_polygon(). Also allow user to specify an export_fname (file name) to save the correspondence table for these new tables as they can’t be named with make_fname().

make create_child_geo() which adapts an sf polygon (child_geo) object by splitting polygons which span across multiple polygons within another passed sf polygon object (parent_geo). Can be used to adapt, for example, SA3s that span across multiple LHN’s so that those that cross are split.
move datasets that are used to hpa.spatial.data pkg and access them just int he same way that strayr reads in data from {absmapsdata}.
properly use export_dir from get_polygon().
remove all internal data and rely only on data from hpa.spatial.data.

use tidy evaluation on quosures to avoid problems with accessing objects passed to hpa.spatial functions that evaluate calls.
save and read correspondence tables to save time particularly when a correspondence table that’s being created is being created many times over within map_data_with_correspondence() (say, when groups is being used).

default mb_geo in map_data_with_correspondence() as missing and apply default within function code. Using a hpa.spatial function as a default arg fails when it’s called by namespace (hpa.spatial::map_data_with_correspondence()).

add get_correspondence_tbl() which calls strayr::read_correspondence_tbl() and if it fails to find one, it uses make_correspondence_tbl() to create one. Takes the same args as strayr::read_correspondence_tbl() but with the option to use a different mesh blocks edition/population data for make_correspondence_tbl().
add make_correspondence_tbl() to create a correspondence table between two given geographies and a POINT geometry (sf) with population counts (Person as is standard in mesh blocks data).
added a .data argument to map_data_with_correspondence() so that codes, values and groups can be passed by reference (like in dplyr::mutate()).
always return a tibble from map_data_with_correspondence() (just for consistency).
add tests for (#3) which is a non-issue.

create an lhn dataset to replace the independent state datasets (qld_hhs, nsw_lhd, sa_lhn etc…).
bug fix (#2): use withr::with_package() and withr::with_environment() to evaluate calls in get_polygon() with hpa.spatial namespace within copy of parent environment.

initial release with two main functions:
- get_polygon() accesses the polygons from {absmapsdata} via strayr::read_absmap() as well as some other shape files internal to this pkg
- map_data_with_correspondence() allows the user to map data (unit or aggregate values) across between editions of the ASGS as well as up within nested levels (i.e. SA2 to SA3).