Getting Started

Introduction

Tidygeocoder provides a unified interface for performing both forward and reverse geocoding queries with a variety of geocoding services. In forward geocoding you provide an address to the geocoding service and you get latitude and longitude coordinates in return. In reverse geocoding you provide the latitude and longitude and the geocoding service will return that location’s address. In both cases, other data about the location can be provided by the geocoding service.

The geocode() and geo() functions are for forward geocoding while the reverse_geocode() and reverse_geo() functions perform reverse geocoding. The geocode() and reverse_geocode() functions extract either addresses (forward geocoding) or coordinates (reverse geocoding) from the input dataframe and pass this data to the geo() and reverse_geo() functions respectively which execute the geocoding queries. All extra arguments (...) given to geocode() are passed to geo() and extra arguments given to reverse_geocode() are passed to reverse_geo().

Forward Geocoding

library(tibble)
library(dplyr)
library(tidygeocoder)

address_single <- tibble(singlelineaddress = c(
  "11 Wall St, NY, NY",
  "600 Peachtree Street NE, Atlanta, Georgia"
))
address_components <- tribble(
  ~street, ~cty, ~st,
  "11 Wall St", "NY", "NY",
  "600 Peachtree Street NE", "Atlanta", "GA"
)

You can use the address argument to specify single-line addresses. Note that when multiple addresses are provided, the batch geocoding functionality of the Census geocoding service is used. Additionally, verbose = TRUE displays logs to the console.

census_s1 <- address_single %>%
  geocode(address = singlelineaddress, method = "census", verbose = TRUE)
#> 
#> Number of Unique Addresses: 2
#> Executing batch geocoding...
#> Batch limit: 10,000
#> Passing 2 addresses to the US Census batch geocoder
#> Querying API URL: https://geocoding.geo.census.gov/geocoder/locations/addressbatch
#> Passing the following parameters to the API:
#> format : "json"
#> benchmark : "Public_AR_Current"
#> vintage : "Current_Current"
#> Query completed in: 1 seconds
singlelineaddress lat long
11 Wall St, NY, NY 40.70747 -74.01122
600 Peachtree Street NE, Atlanta, Georgia 33.77085 -84.38505

Alternatively you can run the same query with the geo() function by passing the address values from the dataframe directly. In either geo() or geocode(), the lat and long arguments are used to name the resulting latitude and longitude fields. Here the method argument is used to specify the “osm” (Nominatim) geocoding service. Refer to the geo() function documentation for the possible values of the method argument.

osm_s1 <- geo(
  address = address_single$singlelineaddress, method = "osm",
  lat = latitude, long = longitude
)
#> Passing 2 addresses to the Nominatim single address geocoder
#> Query completed in: 2 seconds
address latitude longitude
11 Wall St, NY, NY 40.70707 -74.01117
600 Peachtree Street NE, Atlanta, Georgia 33.77086 -84.38614

Instead of single-line addresses, you can use any combination of the following arguments to specify your addresses: street, city, state, county, postalcode, and country.

census_c1 <- address_components %>%
  geocode(street = street, city = cty, state = st, method = "census")
#> Passing 2 addresses to the US Census batch geocoder
#> Query completed in: 2.5 seconds
street cty st lat long
11 Wall St NY NY 40.70747 -74.01122
600 Peachtree Street NE Atlanta GA 33.77085 -84.38505

To return the full geocoding service results (not just latitude and longitude), specify full_results = TRUE. Additionally, for the Census geocoder you can get fields for geographies such as Census tracts by specifying api_options = list(census_return_type = 'geographies'). Be sure to use full_results = TRUE with the “geographies” return type in order to allow the Census geography columns to be returned.

census_full1 <- address_single %>% geocode(
  address = singlelineaddress,
  method = "census", full_results = TRUE, api_options = list(census_return_type = 'geographies')
)
#> Passing 2 addresses to the US Census batch geocoder
#> Query completed in: 1.2 seconds
singlelineaddress lat long id input_address match_indicator match_type matched_address tiger_line_id tiger_side state_fips county_fips census_tract census_block
11 Wall St, NY, NY 40.70747 -74.01122 1 11 Wall St, NY, NY, , , Match Exact 11 WALL ST, NEW YORK, NY, 10005 59659656 R 36 061 000700 1004
600 Peachtree Street NE, Atlanta, Georgia 33.77085 -84.38505 2 600 Peachtree Street NE, Atlanta, Georgia, , , Match Non_Exact 600 PEACHTREE ST, ATLANTA, GA, 30308 17343689 L 13 121 001902 2003

As mentioned earlier, the geocode() function passes addresses in dataframes to the geo() function for geocoding so we can also directly use the geo() function in a similar way:

salz <- geo("Salzburg, Austria", method = "osm", full_results = TRUE) %>%
  select(-licence)
#> Passing 1 address to the Nominatim single address geocoder
#> Query completed in: 1 seconds
address lat long place_id osm_type osm_id boundingbox display_name class type importance icon
Salzburg, Austria 47.79813 13.04648 147539 node 34964314 47.6381346, 47.9581346, 12.8864806, 13.2064806 Salzburg, 5020, Österreich place city 0.6854709 https://nominatim.openstreetmap.org/ui/mapicons//poi_place_city.p.20.png

Reverse Geocoding

For reverse geocoding you’ll use reverse_geocode() instead of geocode() and reverse_geo() instead of geo(). Note that the reverse geocoding functions are structured very similarly to the forward geocoding functions and share many of the same arguments (method, limit, full_results, etc.). For reverse geocoding you will provide latitude and longitude coordinates as inputs and the location’s address will be returned by the geocoding service.

Below, the reverse_geocode() function is used to geocode coordinates in a dataframe. The lat and long arguments specify the columns that contain the latitude and longitude data. The address argument can be used to specify the single line address column name that is returned from the geocoder. Just as with forward geocoding, the method argument is used to specify the geocoding service.

lat_longs1 <- tibble(
  latitude = c(38.895865, 43.6534817),
  longitude = c(-77.0307713, -79.3839347)
)

rev1 <- lat_longs1 %>%
  reverse_geocode(lat = latitude, long = longitude, address = addr, method = "osm")
#> Passing 2 coordinates to the Nominatim single coordinate geocoder
#> Query completed in: 2 seconds
latitude longitude addr
38.89587 -77.03077 Freedom Plaza, 1455, Pennsylvania Avenue Northwest, Washington, District of Columbia, 20004, United States
43.65348 -79.38393 Toronto City Hall, 100, Queen Street West, Financial District, Spadina—Fort York, Old Toronto, Toronto, Golden Horseshoe, Ontario, M5H 2N2, Canada

The same query can also be performed by passing the latitude and longitudes directly to the reverse_geo() function. Here we will use full_results = TRUE so that the full results are returned (not just the single line address column).

rev2 <- reverse_geo(
  lat = lat_longs1$latitude,
  long = lat_longs1$longitude,
  method = "osm",
  full_results = TRUE
)
#> Passing 2 coordinates to the Nominatim single coordinate geocoder
#> Query completed in: 2 seconds
glimpse(rev2)
#> Rows: 2
#> Columns: 22
#> $ lat            <dbl> 38.89587, 43.65348
#> $ long           <dbl> -77.03077, -79.38393
#> $ address        <chr> "Freedom Plaza, 1455, Pennsylvania Avenue Northwest, Washington, District of Columbia, 20004, United States", "Toronto City Hall, 1…
#> $ place_id       <int> 284009208, 148364261
#> $ licence        <chr> "Data © OpenStreetMap contributors, ODbL 1.0. https://osm.org/copyright", "Data © OpenStreetMap contributors, ODbL 1.0. https://osm…
#> $ osm_type       <chr> "relation", "way"
#> $ osm_id         <int> 8060882, 198500761
#> $ osm_lat        <chr> "38.895849999999996", "43.6536032"
#> $ osm_lon        <chr> "-77.03077367444483", "-79.38400546703345"
#> $ tourism        <chr> "Freedom Plaza", NA
#> $ house_number   <chr> "1455", "100"
#> $ road           <chr> "Pennsylvania Avenue Northwest", "Queen Street West"
#> $ city           <chr> "Washington", "Old Toronto"
#> $ state          <chr> "District of Columbia", "Ontario"
#> $ postcode       <chr> "20004", "M5H 2N2"
#> $ country        <chr> "United States", "Canada"
#> $ country_code   <chr> "us", "ca"
#> $ boundingbox    <list> <"38.8956276", "38.896068", "-77.03182", "-77.0297273">, <"43.6529946", "43.6541458", "-79.3848438", "-79.3830415">
#> $ amenity        <chr> NA, "Toronto City Hall"
#> $ neighbourhood  <chr> NA, "Financial District"
#> $ quarter        <chr> NA, "Spadina—Fort York"
#> $ state_district <chr> NA, "Golden Horseshoe"

Working With Messy Data

Only unique input data (either addresses or coordinates) is passed to geocoding services even if your data contains duplicates. NA and blank inputs are excluded from queries. Input latitudes and longitudes are also limited to the range of possible values.

Below is an example of how duplicate and missing data is handled by tidygeocoder. As the console messages shows, only the two unique addresses are passed to the geocoding service.

# create a dataset with duplicate and NA addresses
duplicate_addrs <- address_single %>%
  bind_rows(address_single) %>%
  bind_rows(tibble(singlelineaddress = rep(NA, 3)))

duplicates_geocoded <- duplicate_addrs %>%
  geocode(singlelineaddress, verbose = TRUE)
#> 
#> Number of Unique Addresses: 2
#> Passing 2 addresses to the Nominatim single address geocoder
#> 
#> Number of Unique Addresses: 1
#> Querying API URL: https://nominatim.openstreetmap.org/search
#> Passing the following parameters to the API:
#> q : "11 Wall St, NY, NY"
#> limit : "1"
#> format : "json"
#> HTTP Status Code: 200
#> Query completed in: 0.2 seconds
#> Total query time (including sleep): 1 seconds
#> 
#> 
#> Number of Unique Addresses: 1
#> Querying API URL: https://nominatim.openstreetmap.org/search
#> Passing the following parameters to the API:
#> q : "600 Peachtree Street NE, Atlanta, Georgia"
#> limit : "1"
#> format : "json"
#> HTTP Status Code: 200
#> Query completed in: 0.2 seconds
#> Total query time (including sleep): 1 seconds
#> 
#> Query completed in: 2 seconds
singlelineaddress lat long
11 Wall St, NY, NY 40.70707 -74.01117
600 Peachtree Street NE, Atlanta, Georgia 33.77086 -84.38614
11 Wall St, NY, NY 40.70707 -74.01117
600 Peachtree Street NE, Atlanta, Georgia 33.77086 -84.38614
NA NA NA
NA NA NA
NA NA NA

As shown above, duplicates will not be removed from your results by default. However, you can return only unique results by using unique_only = TRUE. Note that passing unique_only = TRUE to geocode() or reverse_geocode() will result in the original dataframe format (including column names) to be discarded in favor of the standard field names (ie. “address”, ‘lat, ’long’, etc.).

uniqueonly1 <- duplicate_addrs %>%
  geocode(singlelineaddress, unique_only = TRUE)
#> Passing 2 addresses to the Nominatim single address geocoder
#> Query completed in: 2 seconds
address lat long
11 Wall St, NY, NY 40.70707 -74.01117
600 Peachtree Street NE, Atlanta, Georgia 33.77086 -84.38614

Combining Multiple Queries

The geocode_combine() function allows you to execute and combine the results of multiple geocoding queries. The queries are specified as a list of lists with the queries parameter and are executed in the order provided. By default only addresses that are not found are passed to the next query, but this behavior can be toggled with the cascade argument.

In the first example below, the US Census service is used for the first query while the Nominatim (“osm”) service is used for the second query. The global_params argument passes the address column from the input dataset to both queries.

addresses_combine <- tibble(
  address = c('100 Wall Street NY, NY', 'Paris', 'Not An Address')
)

cascade_results1 <- addresses_combine %>%
  geocode_combine(
    queries = list(
      list(method = 'census'),
      list(method = 'osm')
    ),
    global_params = list(address = 'address')
  )
#> 
#> Passing 3 addresses to the US Census batch geocoder
#> Query completed in: 0.2 seconds
#> Passing 2 addresses to the Nominatim single address geocoder
#> Query completed in: 2 seconds
address lat long query
100 Wall Street NY, NY 40.70516 -74.007350 census
Paris 48.85889 2.320041 osm
Not An Address NA NA

If cascade is set to FALSE then all addresses are attempted by each query regardless of if the address was found initially or not.

no_cascade_results1 <- addresses_combine %>%
  geocode_combine(
    queries = list(
      list(method = 'census'),
      list(method = 'osm')
    ),
    global_params = list(address = 'address'),
    cascade = FALSE
  )
#> 
#> Passing 3 addresses to the US Census batch geocoder
#> Query completed in: 0.3 seconds
#> Passing 3 addresses to the Nominatim single address geocoder
#> Query completed in: 3 seconds
address lat long query
100 Wall Street NY, NY 40.70516 -74.007350 census
100 Wall Street NY, NY 40.70522 -74.006800 osm
Paris NA NA census
Paris 48.85889 2.320041 osm
Not An Address NA NA census
Not An Address NA NA osm

Additionally, the results from each query can be returned in separate list items by setting return_list = TRUE.

Customizing Queries

The limit argument can be specified to allow multiple results (rows) per input if available. The maximum value for the limit argument is often 100 for geocoding services. To use the default limit value for the selected geocoding service you can use limit = NULL which will prevent the limit parameter from being included in the query.

geo_limit <- geo(
  c("Lima, Peru", "Cairo, Egypt"),
  method = "osm",
  limit = 3, full_results = TRUE
)
#> Passing 2 addresses to the Nominatim single address geocoder
#> Query completed in: 2 seconds
glimpse(geo_limit)
#> Rows: 6
#> Columns: 13
#> $ address      <chr> "Lima, Peru", "Lima, Peru", "Lima, Peru", "Cairo, Egypt", "Cairo, Egypt", "Cairo, Egypt"
#> $ lat          <dbl> -11.96784, -12.03089, -11.95785, 30.04439, 30.08695, 30.22503
#> $ long         <dbl> -77.01094, -77.09072, -77.04139, 31.23573, 31.96162, 31.69733
#> $ place_id     <int> 128601425, 118029294, 128061402, 283020077, 216381765, 272362266
#> $ licence      <chr> "Data © OpenStreetMap contributors, ODbL 1.0. https://osm.org/copyright", "Data © OpenStreetMap contributors, ODbL 1.0. https://osm.o…
#> $ osm_type     <chr> "way", "way", "way", "relation", "way", "way"
#> $ osm_id       <int> 116948976, 71187508, 115715296, 5466227, 544272017, 914722135
#> $ boundingbox  <list> <"-11.9678367", "-11.9672995", "-77.0117952", "-77.0109387">, <"-12.0308994", "-12.0304024", "-77.0911106", "-77.090276">, <"-11.9608…
#> $ display_name <chr> "Peru, San Juan de Lurigancho, Huascar, Lima, Lima Metropolitana, Lima, 15423, Perú", "Peru, San Martín de Porres, Lima, Lima Metropo…
#> $ class        <chr> "highway", "highway", "highway", "place", "highway", "highway"
#> $ type         <chr> "residential", "residential", "residential", "city", "motorway", "trunk"
#> $ importance   <dbl> 0.3200000, 0.3200000, 0.3200000, 0.6960286, 0.1000000, 0.1000000
#> $ icon         <chr> NA, NA, NA, "https://nominatim.openstreetmap.org/ui/mapicons//poi_place_city.p.20.png", NA, NA

To directly specify specific API parameters for a given method you can use the custom_query parameter. For example, the Nominatim (OSM) geocoder has a ‘polygon_geojson’ argument that can be used to return GeoJSON geometry content. To pass this parameter you can insert it with a named list using the custom_query argument:

cairo_geo <- geo("Cairo, Egypt",
  method = "osm", full_results = TRUE,
  custom_query = list(polygon_geojson = 1), verbose = TRUE
)
#> 
#> Number of Unique Addresses: 1
#> Passing 1 address to the Nominatim single address geocoder
#> 
#> Number of Unique Addresses: 1
#> Querying API URL: https://nominatim.openstreetmap.org/search
#> Passing the following parameters to the API:
#> q : "Cairo, Egypt"
#> limit : "1"
#> polygon_geojson : "1"
#> format : "json"
#> HTTP Status Code: 200
#> Query completed in: 0.2 seconds
#> Total query time (including sleep): 1 seconds
#> 
#> Query completed in: 1 seconds
glimpse(cairo_geo)
#> Rows: 1
#> Columns: 15
#> $ address             <chr> "Cairo, Egypt"
#> $ lat                 <dbl> 30.04439
#> $ long                <dbl> 31.23573
#> $ place_id            <int> 283020077
#> $ licence             <chr> "Data © OpenStreetMap contributors, ODbL 1.0. https://osm.org/copyright"
#> $ osm_type            <chr> "relation"
#> $ osm_id              <int> 5466227
#> $ boundingbox         <list> <"29.7483062", "30.3209168", "31.2200331", "31.9090054">
#> $ display_name        <chr> "القاهرة, محافظة القاهرة, مصر"
#> $ class               <chr> "place"
#> $ type                <chr> "city"
#> $ importance          <dbl> 0.6960286
#> $ icon                <chr> "https://nominatim.openstreetmap.org/ui/mapicons//poi_place_city.p.20.png"
#> $ geojson.type        <chr> "Polygon"
#> $ geojson.coordinates <list> <<array[1 x 119 x 2]>>

To test a query without sending any data to a geocoding service, you can use no_query = TRUE (NA results are returned).

noquery1 <- geo(c("Vancouver, Canada", "Las Vegas, NV"),
  no_query = TRUE,
  method = "arcgis"
)
#> 
#> Number of Unique Addresses: 2
#> Passing 2 addresses to the ArcGIS single address geocoder
#> 
#> Number of Unique Addresses: 1
#> Querying API URL: https://geocode.arcgis.com/arcgis/rest/services/World/GeocodeServer/findAddressCandidates
#> Passing the following parameters to the API:
#> SingleLine : "Vancouver, Canada"
#> maxLocations : "1"
#> f : "json"
#> 
#> Number of Unique Addresses: 1
#> Querying API URL: https://geocode.arcgis.com/arcgis/rest/services/World/GeocodeServer/findAddressCandidates
#> Passing the following parameters to the API:
#> SingleLine : "Las Vegas, NV"
#> maxLocations : "1"
#> f : "json"
#> Query completed in: 0 seconds
address lat long
Vancouver, Canada NA NA
Las Vegas, NV NA NA

Additional usage notes for the geocode(), geo(), reverse_geocode(), and reverse_geo() functions: