Skip to main content

R for Audit Analytics

38 Reverse geo-coding

38.1 Introduction

Geo-coding is the process of converting a human-readable address into geographic coordinates, such as latitude and longitude. Geocoding is the process of determining the location of an address on a map. So, plotting the places on geographical map layers require extraction of latitudes and longitudes from master databases.

Reverse geo-coding, on the other hand, is the process of converting geographic coordinates (latitude and longitude) into a human-readable address, such as a street address, city, state/province, and country. In other words, it’s the process of determining a location’s address based on its geographic coordinates.

Both geo-coding and reverse geo-coding are important in geospatial data analysis, as they allow us to link spatial information with non-spatial information, such as demographic data, land use data, or other forms of spatial data. This information can then be used to gain insights into patterns and relationships in the data, and to make informed decisions based on location information.

Reverse geo-coding can be used in forensic data analysis to help identify the location of transactions or events. For example, it can be used to determine the location of an ATM or point of sale device where a fraudulent transaction occurred. It can also be used to analyze patterns of activity in a particular geographic area, such as the frequency and timing of certain types of transactions. As another example, we may think of an application data, where mobile or hand held devices are used to capture data at the time and place of intended service delivery. Auditors may use the geo-spatial data captured by these devices to cross check/verify the actual points of service delivery.

38.2 Widely used databases

OpenStreetMap (OSM) is one of the most popular and widely used master databases for geo-coding and reverse geo-coding. OSM is a free and open-source map of the world, created and maintained by a community of volunteers. It provides a rich and detailed set of spatial data, including street names, addresses, and other points of interest.

Other popular master databases for geo-coding and reverse geo-coding include Google Maps. Google Maps provides a rich set of geo-coding and reverse geo-coding APIs that are widely used in web and mobile applications. Google Maps provides a comprehensive set of spatial data, including street names, addresses, and other points of interest.

38.3 Example in R

38.3.1 Prerequisites

38.3.2 Function to be used

We will use reverse_geo() function from this library. Syntax is

reverse_geo(
  lat,
  long,
  method = "osm",
  address = "address",
  full_results = FALSE
  ...
)

Here -

  • lat and long are, as the names suggest, latitude and longitudes for the place, of which address is to be extracted.
  • method argument provides for geo service to be used. Default is osm. Other services available are - arcgis, geocodio, google, etc.
  • address provides the column name to be used.
  • full_results by default is FALSE. But if set to TRUE returns all available data from the geocoding service.
  • for other arguments, please check help of the function.

38.3.3 Practical data-set

Remember one thing, the function is not vectorised. So we will have to use apply family or purrr::map family of functions to get reverse geo-coding information.

# Coordinates of tajmahal
reverse_geo(lat = 27.1751, 
            long = 78.421, 
            full_results = TRUE, 
            method = 'osm')
## # A tibble: 1 × 25
##     lat  long address     place_id licence osm_type osm_id osm_lat osm_lon class
##   <dbl> <dbl> <chr>          <int> <chr>   <chr>     <int> <chr>   <chr>   <chr>
## 1  27.2  78.4 Sailai, Fi…   2.27e8 Data ©… way      4.56e8 27.175… 78.421… high…
## # ℹ 15 more variables: type <chr>, place_rank <int>, importance <dbl>,
## #   addresstype <chr>, name <chr>, suburb <chr>, town <chr>, county <chr>,
## #   state_district <chr>, state <chr>, `ISO3166-2-lvl4` <chr>, postcode <chr>,
## #   country <chr>, country_code <chr>, boundingbox <list>

38.3.4 Using it in a dataset

Let’s build a sample dataset of say 4 values/places of interest in India

lat long
27.1751 78.0421
24.8318 79.9199
27.1795 78.0211
28.5245 77.1855

Now extracting information for these using purrr::map family

library(dplyr, warn.conflicts = FALSE)
library(purrr, warn.conflicts = FALSE)

results <- pmap_dfr(sample,
     ~ reverse_geo(..1, # first column in sample
                   ..2, # second column in sample
                   full_results = TRUE)) %>% 
  select(lat, long, country, state, city, village, address, postcode)

38.3.5 View above results

lat long country state city village address postcode
27.1751 78.0421 India Uttar Pradesh Agra NA Taj Mahal, Taj East Gate Road, Taj Ganj, Agra, Uttar Pradesh, 282004, India 282004
24.8318 79.9199 India Madhya Pradesh NA Jatkra Khajuraho, Jatkra, Rajnagar Tahsil, Chhatarpur, Madhya Pradesh, 471006, India 471006
27.1795 78.0211 India Uttar Pradesh Agra NA Agra Fort, Minar Bazaar, Subhash Bazaar, Agra, Uttar Pradesh, 280001, India 280001
28.5245 77.1855 India Delhi NA NA Qutub Minar Complex, Kalka Das Marg, Mehrauli, Mehrauli Tehsil, South Delhi, Delhi, 110030, India 110030

38.3.6 Alternate syntax (for those who do not want to use purrr::pmap_dfr) 

sapply(seq(nrow(sample)),
       function(x) reverse_geo(sample$lat[x], sample$long[x], full_results = TRUE)) %>% 
  map_dfr(rbind)