Here I will introduce you to the osmdata package. osm is short for open street map and you can use the package to access the data from osm directly from R.

install.packages("osmdata")
libary(osmdata)
library(sf)
library(dplyr)
library(tmap)
library(magrittr)

osm uses the read-only overpass API to download data from Open Street Map and read them in the formats of sp, sf or silicate. A complete call with osmdata looks like his:

x <-
  getbb("Landau") %>% 
        opq() %>%
        add_osm_feature(key = 'name',
                        value = "Queich") %>%
        osmdata_sf()

As you can see, four different functions are part of the call.

  1. getbb()
  2. opq()
  3. add_osm_feature()
  4. osmdata_sf()

Lets go through them one after another to see what they do. getbb() returns the bounding box of some place in the world. The only argument you usually have to supply is place_name.

ld = getbb("Landau")

This worked, and returned the bounding box of Landau in longitude and latitude. Using our new bounding box we can now call opq(). Of course you can also write a custom bbox directly into opq(). This function builds a query to Overpass (overpass query).

ld = opq(bbox = ld)
ld
## $bbox
## [1] "49.1548639,7.8646346,49.3163444,8.1806097"
## 
## $prefix
## [1] "[out:xml][timeout:25];\n(\n"
## 
## $suffix
## [1] ");\n(._;>;);\nout body;"
## 
## $features
## NULL
## 
## attr(,"class")
## [1] "list"           "overpass_query"
## attr(,"nodes_only")
## [1] FALSE

This is no spatial object yet but just the request (query) that will be send to the API. This request is further expanded in add_osm_feature(). All items in the Open Street Map data have a key and a value. A few example are key: amenity - value: bar, cafe or fast_food; key: building - value: school, public or church, key: landuse - value: basin, cementery, grass. See here for a complete overview of all keys and features. Let’s download one of each.

ld_amenity = add_osm_feature(ld, key = "amenity", value = "bar") 
ld_building = add_osm_feature(ld, key = "building", value = "school") 
ld_landuse = add_osm_feature(ld, key = "landuse", value = "grass")

Now our queries are complete and we can download the data using osmdata_sf().

ld_amenity  %<>% osmdata_sf() 
## Error in curl::curl_fetch_memory(url, handle = handle): Timeout was reached: [overpass-api.de] Operation timed out after 10000 milliseconds with 0 out of 0 bytes received
## Request failed [ERROR]. Retrying in 1.6 seconds...
ld_building %<>% osmdata_sf() 
ld_landuse  %<>% osmdata_sf()

The suffix of osmdata_sf() specifies that we want the downloaded data to be in the sf format. Alternatives are: osmdata_sp,osmdata_sc and osmdata_xml.

The resulting objects are lists which have the different geometries as elements. Below we plot the three data sets together.

tmap_mode("view")
## tmap mode set to interactive viewing
tm_shape(ld_amenity$osm_points) + tm_dots(col = "red") + 
  tm_shape(ld_building$osm_polygons) + tm_polygons(col = "blue") + 
  tm_shape(ld_landuse$osm_polygons) + tm_polygons(col = "green")