In etiennebr/geotidy: Tidy manipulation of spatial data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

geotidy

Manipulate spatial data tidily. Geotidy provides a selection of spatial functions from the sf package that are adapted to a tidy workflow. It relies on tibbles and dplyr verbs, and forces you to be explicit about the spatial operations you desire. The package is experimental. If you want to learn more about the motivations behind it, see the [Motivation] section.

Installation

You can install the experimental version of geotidy from github :

# install.packages("remotes")
remotes::install_github("etiennebr/geotidy")

Example

Here's how you manipulate spatial data with geotidy.

library(tibble)
library(dplyr)
library(geotidy)

tibble(place = "Sunset Room", longitude = -97.7404985, latitude = 30.2645315) %>% 
  mutate(geometry = st_point(longitude, latitude))

Motivation

Many spatial formats use a model where a geometry can have many attributes. This model can be very useful in many applications, but it sometimes conflicts with the principles of tidy data, where one observation is one row. To keep spatial data and operations tidy, geotidy does less than sf. It makes manipulations explicit by forcing the use of a function on the geometry column. It doesn't try to guess which is the geometry column that should receive the operation. It also makes it clear by reading the code, which geometry is impacted. This is done by treating geometry columns just like other tibble columns. sf often hides the geometry column, geotidy treats it just like a regular columns. This also makes it easier to interact with other OGC compliant tools, such as postgis or spark+geomesa.

Example

While sf will guess which column should be buffered: ``` {r eval = FALSE} shp <- sf::st_read("") st_buffer(shp)

`geotidy` forces to be explicit and use `dplyr` verbs
```r
shp %>% 
  mutate(geometry = st_buffer(geometry))

If you already use dplyr with sf, geotidy should feel natural and remove some of the casting operations. geotidy guarantees that your data will stay tidy from start to finish. By having explicit management of geometry columns, it is also easy to track multiple columns.

geotidy also guarantees that the returned values are either scalar, or a vector or a list with the same length than the original geometry and not drop any data without the user consent (looking at you st_cast!).

geotidy is not a fork or a separation from sf. It just adds a constrained layer on top of sf to facilitate a tidy workflow. It is an experiment that could be integrated in sf and is likely to change.

etiennebr/geotidy documentation built on Feb. 28, 2020, 4:45 a.m.