gissr

gissr is a collection of R functions which make working with spatial data easier.

Retired note

As most spatial data and R users will be aware, the rgdal, rgeos, and maptools packages were retired in October 2023. This is because the developer of the packages retired and a new generation of spatial tools has emerged in the form of the sf, terra, and stars packages. Because gissr is mostly built upon the older packages (that have now been archived), gissr was retired on June 17, 2024. The development of gissr's successor based on the sf and terra packages called sspatialr is ongoing. For new projects, it is recommended that sspatialr is used rather than gissr.

Installation

The development version:

# Install dependency
remotes::install_github("skgrange/threadr")

# Install gissr
remotes::install_github("skgrange/gissr")

Background

R's spatial data analysis abilities are very well developed. Therefore, R can be an effective geographical information system (GIS). A key advantage of R in GIS applications is that the user can dip in and out of R's general string, numerical, and visualisation tools and apply them to spatial data.

However, the challenges I have with using R as a GIS are:

• Keeping track of the multiple packages which are used,

• the lack of consistency among these packages, and

• the lack of tidy outputs which other areas of the R ecosystem have been so good at developing.

To overcome these points, I have written wrappers for many geographical functions which generally begin sp_ to do particular tasks and bundle all the dependencies together as a package. Some of these functions will likely be useful for others.

Utility functions

• Easily read shapefiles, GPX, GeoJSON, KML, GML, TAB, and File Geodatabases with sp_read, a wrapper for rgdal::readOGR.

  • Also check spatial files and system things with sp_list_drivers, sp_list_layers, and sp_layer_info.

• Transform projection systems with sp_transform.

  • sp_transform can also force projections when a spatial object has none.
  • transform_coordinates does a similar thing, but for data frames.

• Transform data frames (tables) to spatial points, lines, or polygons with sp_from_data_frame.

• Transform data frames with a well-known text (WKT) variable (or just a vector) to a spatial object with sp_from_wkt.

• Bind/combine different spatial objects with sp_bind.

• Unite spatial objects with sp_unite and do the opposite with sp_disaggregate.

• Calculate lengths or areas of spatial objects with sp_area and sp_length.

• Clip or crop a spatial object to a rectangular envelope with sp_clip.

  • To filter objects by other polygons, use [ subsetting (or sp_filter).

  • Rectangular or elliptical polygons can be created with sp_create_envelope and sp_ellipse for this purpose too.

• Do simple transformations to spatial objects with sp_move, sp_flip, sp_reflect, and sp_rotate.

• Simplify spatial objects with sp_simplify.

• "Dissolve" polygons to make a single feature with sp_dissolve_polygons.

• "Punch" holes in polygons with sp_punch.

• Add positive or negative buffers with sp_buffer.

• Create enclosing polygons with sp_convex_hull.

• Find centroids of geometries with sp_centroid.

• "Promote" or "demote" Spatial* to Spatial*DataFrame, i.e. add or drop data slots for geometries with sp_promote and sp_demote.

• Return and reset geometry IDs with sp_feature_ids and sp_reset_feature_ids.

• Point-in-polygon tests with sp_left_join.

• Calculate distances among spatial objects with sp_distance.

  • distance_by_haversine does the same thing, but with a different method, and for data frames.

• Fix issues with spatial objects with sp_fix. This function is a blatant wrap of cleangeo::clgeo_Clean. This function is a good piece of work so make sure you have a look at the cleangeo package.

• Parse vectors of degrees, minutes, and seconds into decimal degrees with dms_to_decimal.

• Sort/arrange points in a clockwise order with sort_points.

• Create Tessellation polygons with sp_tessellation_polygons.

• Export spatial objects to spatial data files with write_gpx, write_geojson, and write_shapefile.

• Transform spatial objects to data frames with sp_fortify.

Raster functions

• Create a raster layer from spatial data with ra_from_sp.

• Filter/crop/mask a raster layer with a spatial polygon with ra_mask.

• Interpolate a raster layer/surface with ra_interpolate.

• Increase a raster's resolution with ra_disaggregate.

• Smooth a raster's values with ra_focal.

• Extract values from raster objects using spatial data types with ra_drill and then produce a "tidy data" version with tidy_ra_drill.

• Transform raster objects to data frames with ra_fortify.

• Bind or merge a number of raster layers together with ra_bind

OpenStreetMap data importers

• A collection of get_osm_* functions to import data from OpenStreetMap.

Things I want to do

• Develop a function which can read n features in a spatial data file. This will be helpful when large data files are encountered and system memory is too small to load the entire file at once.

• Get the interface between R and SpatiaLite sorted -- this can probably be left to sf now.

• Concave hull function i.e. find minimum area polygon.

• Add support for WKB (well-known binary).

See also

• sf
• terra
• sspatialr