Slides: https://nowosad.github.io/SIGR2021/workshop1/workshop1_rl.html#1

Example datasets

Why use example Datasets

learning and documenting the use of the package
use a small dataset rather than a large dataset => quicker
unit testing

size of a sample dataset: - small enough to be fast - but with edge cases

You can have multiple datasets (small/large, differents areas)

Robin recommends to create an example dataset to save time. Documenting it is recommended but not mandatory.

ActDev example: know where to stop R: R is the backend, the webapplication to visualise the data is in JS.

library(help ="datasets") # full list of datasets

Listed datasets are available directly from R.

Plot 2 datasets

plot(trees)

plot(LakeHuron)

Get information on data in a package

help(package = "spData")

?AirPassengers

library(spData)
plot(nz) # if list error, maybe because sf is not loaded yet and plot don't know how to plot sf objects
#> Error in data.matrix(x): 'list' object cannot be coerced to type 'double'

plot(nz$geom)
#> Error in xy.coords(x, y, xlabel, ylabel, log): 'x' is a list, but does not have components 'x' and 'y'

Make a copy to work on is good practice.

nz2 = nz
plot(nz)
#> Error in data.matrix(x): 'list' object cannot be coerced to type 'double'

library(nzelect)

nz_lonlat = sf::st_transform(nz, 4326)
names(voting_places)
#>  [1] "electorate_number"   "electorate"          "voting_place_id"    
#>  [4] "voting_place_suburb" "northing"            "easting"            
#>  [7] "longitude"           "latitude"            "voting_place"       
#> [10] "election_year"       "coordinate_system"   "TA2014_NAM"         
#> [13] "REGC2014_N"          "AU2014"              "MB2014"

# create points from coordinates
voting_places_sf = sf::st_as_sf(voting_places, coords = c("longitude", "latitude"))

plot(sf::st_geometry(nz_lonlat))
plot(voting_places_sf, add = TRUE)
#> Warning in plot.sf(voting_places_sf, add = TRUE): ignoring all but the first
#> attribute

Spatial subsetting

Filtering voting places that are inside NZ island

vsub = voting_places_sf[nz,] # Error message because mismatching CRS issues
#> Error in st_geos_binop("intersects", x, y, sparse = sparse, prepared = prepared, : st_crs(x) == st_crs(y) is not TRUE

sf::st_crs(voting_places_sf, error=TRUE) # NA
#> Coordinate Reference System: NA
sf::st_crs(nz) # EPSG:2193
#> Coordinate Reference System:
#>   User input: EPSG:2193 
#>   wkt:
#> PROJCS["NZGD2000 / New Zealand Transverse Mercator 2000",
#>     GEOGCS["NZGD2000",
#>         DATUM["New_Zealand_Geodetic_Datum_2000",
#>             SPHEROID["GRS 1980",6378137,298.257222101,
#>                 AUTHORITY["EPSG","7019"]],
#>             TOWGS84[0,0,0,0,0,0,0],
#>             AUTHORITY["EPSG","6167"]],
#>         PRIMEM["Greenwich",0,
#>             AUTHORITY["EPSG","8901"]],
#>         UNIT["degree",0.0174532925199433,
#>             AUTHORITY["EPSG","9122"]],
#>         AUTHORITY["EPSG","4167"]],
#>     PROJECTION["Transverse_Mercator"],
#>     PARAMETER["latitude_of_origin",0],
#>     PARAMETER["central_meridian",173],
#>     PARAMETER["scale_factor",0.9996],
#>     PARAMETER["false_easting",1600000],
#>     PARAMETER["false_northing",10000000],
#>     UNIT["metre",1,
#>         AUTHORITY["EPSG","9001"]],
#>     AUTHORITY["EPSG","2193"]]

Voting_places_sf does not have a crs

Fixing issues

voting_places_sf <- sf::st_set_crs(voting_places_sf, st_crs(4326))
#> Error in st_crs(4326): could not find function "st_crs"
sf::st_crs(voting_places_sf)
#> Coordinate Reference System: NA
nz_lonlat <- sf::st_transform(crs = st_crs(4326))
#> Error in st_crs(4326): could not find function "st_crs"
vsub = voting_places_sf[nz_lonlat,]
#> Error in st_geos_binop("intersects", x, y, sparse = sparse, prepared = prepared, : st_crs(x) == st_crs(y) is not TRUE

plot(sf::st_geometry(nz_lonlat))

plot(vsub, add = TRUE)
#> Error in plot(vsub, add = TRUE): object 'vsub' not found

Using OSM data

Quite informal but very very rich Interesting when you can’t access official data

Simple requests: {osmdata}

library(osmdata)
#> Error in library(osmdata): there is no package called 'osmdata'

schools_nz_osm <- opq(bbox = sf::st_bbox(nz_lonlat)) %>%
  add_osm_feature(key = "amenity", value = "school")
#> Error in opq(bbox = sf::st_bbox(nz_lonlat)) %>% add_osm_feature(key = "amenity", : could not find function "%>%"

Working with messy data

skimr:skim : resume data

selecting variables of interest : saving space and time

source outside code

source("https://raw.githubusercontent.com/Nowosad/SIGR2021/master/workshop1/osm_combine.R")

# osm_combine() is available in the environnment

Big OSM data

For large dataset (> 100 MB)

library(osmextract)
#> Error in library(osmextract): there is no package called 'osmextract'

Needs a place name, and parameters (date type (e.g. “amenity”) and an SQL query for the OSM server)

Tidyverse

library(spData)
library(sf)
#> Linking to GEOS 3.9.1, GDAL 3.3.0, PROJ 8.0.1
library(tidyverse)
#> Error in library(tidyverse): there is no package called 'tidyverse'

nz %>%
  filter(Island == "South")
#> Error in data.matrix(data): 'list' object cannot be coerced to type 'double'

Geometry operations

Simplification : rmapshaper::ms_simplify
st_is_valid() : check if geometries are valid
st_make_valid() : fixes geometries

download online data

download.file
readr can read some ressources ##

GIS workshop

Jakub Nowosad

Robin Lovelace