pgGeocoder - Japanese Geocoder on PostgreSQL

Requirements

• PostgreSQL >= 11
• PostGIS >= 2.5

For Address Data Import and Creation

• GDAL >= 2.1
• curl
• unzip
• iconv

Setup

1. Copy .env.example to .env.

   $ cd /path/to/pgGeocoder
   $ cp .env.example .env

2. Open .env file with some text editor and adjust database settings and the years of the dataset to import.

   DBROLE=postgres
   DBPASS=postgres
   DBNAME=addresses
   DBHOST=localhost
   DBPORT=5432
   YEAR_ISJ=2023
   YEAR_KSJ=2023
   YEAR_ESTAT=2020
   

   Note: To download and install address data from multiple years to have a historical data set, the YEAR_ISJ parameter can be set as follows:

   YEAR_ISJ="2023,2017,2013"
   

   Be aware though that this will require substantial disk space resources.

3. Create address database (with same as .env values).
   (If the database exists, drop it at first.)

   $ dropdb -U postgres addresses
   $ createdb -U postgres addresses

4. Run install and download/import scripts.

   $ bash scripts/install.sh
   $ bash scripts/import_data.sh

5. Run maintenance script.

   $ bash scripts/maintenance.sh

   If keeping original data source tables is preferable, add 1 as the argument.

   $ bash scripts/maintenance.sh 1

   Note: The maintainance script will take around 45 mins to complete depending on the specs of the computer.

Tables/Functions

About tables structure, check the following files.

• createTables.sql: Define each tables and columns, but without indexes.
• maintTables.sql: Define indexes.

About functions, check the following files.

• pgGeocoder.sql: Define geocoder function.
• pgReverseGeocoder.sql: Define reverse_geocoder function.

Examples

$ psql -U postgres addresses

• Geocode on address:

  select * from geocoder('京都府京都市中京区河原町通四条上る米屋町３８０－１ツジクラビル１階');

   code |     x      |     y     |          address          | todofuken | shikuchoson | ooaza | chiban | go 
  ------+------------+-----------+---------------------------+-----------+-------------+-------+--------+----
      2 | 135.769661 | 35.004476 | 京都府京都市中京区米屋町380番 | 京都府     | 京都市中京区  | 米屋町 | 380    | 
  (1 row)
  

  select * from geocoder('神奈川県横浜市西区みなとみらい３−６−３');

   code |     x      |     y     |             address             | todofuken | shikuchoson |     ooaza      | chiban | go 
  ------+------------+-----------+---------------------------------+-----------+-------------+----------------+--------+----
      1 | 139.632805 | 35.458282 | 神奈川県横浜市西区みなとみらい三丁目6番 | 神奈川県   | 横浜市西区    | みなとみらい三丁目 | 6      | 3
  (1 row)
  

• Reverse geocode on address:

  select * from reverse_geocoder(141.342094, 43.050264);

   code |         x         |         y         |                address                | todofuken | shikuchoson  |      ooaza       | chiban | go 
  ------+-------------------+-------------------+---------------------------------------+-----------+--------------+------------------+--------+----
      1 | 141.3421173095703 | 43.05035400390625 | 北海道札幌市中央区南七条西十一丁目4-5 | 北海道    | 札幌市中央区 | 南七条西十一丁目 | 4      | 5
   (1 row)
  

• Reverse geocode a coordinate and specify search distance in meters (lon, lat, meters)

  select * from reverse_geocoder(141.342094, 43.050264, 50);

   code |         x         |         y         |                address                | todofuken | shikuchoson  |      ooaza       | chiban | go 
  ------+-------------------+-------------------+---------------------------------------+-----------+--------------+------------------+--------+----
      1 | 141.3421173095703 | 43.05035400390625 | 北海道札幌市中央区南七条西十一丁目4-5 | 北海道    | 札幌市中央区 | 南七条西十一丁目 | 4      | 5
   (1 row)
  

Data Sources

1. 位置参照情報 (ISJ)
   • Website: http://nlftp.mlit.go.jp/isj/index.html
   • Format: CSV (Zipped)
   • Geometry Type: Point
   • Remarks:
     • Point based address data for "Gaiku Level" (街区レベル) and "Oaza Level" (大字・町丁目レベル).
2. e-Stat 国勢調査町丁・字等別境界データ
   • Website: https://www.e-stat.go.jp/gis/statmap-search?page=1&type=2&aggregateUnitForBoundary=A&toukeiCode=00200521
   • Format: ESRI Shapefile (or GML)
   • Geometry Type: Polygon
   • Remarks:
     • Almost "Oaza Level" (大字・町丁目レベル) admin boundary data, but some boundaries are merged for Japanese census survey units.
     • Each prefectures' boundaries are not adjusted (snapped), so some overlaps and gaps exist.
3. 国土数値情報 (KSJ)
   • Website: https://nlftp.mlit.go.jp/ksj/index.html
   • Format: ESRI Shapefile (or GML)
   • 行政区域データ:
     • Website: https://nlftp.mlit.go.jp/ksj/gml/datalist/KsjTmplt-N03-2023.html
     • Geometry Type: Polygon
     • Remarks:
       • "City Level" (市区町村レベル) admin boundary data.
   • 市区町村役場データ:
     • Website: https://nlftp.mlit.go.jp/ksj/gml/datalist/KsjTmplt-P34.html
     • Geometry Type: Point
     • Remarks:
       • "City Office" (市区町村役場) point data.
   • 国・都道府県の機関データ:
     • Website: https://nlftp.mlit.go.jp/ksj/gml/datalist/KsjTmplt-P28-2022.html
     • Geometry Type: Point
     • Remarks:
       • Geovernment data which includes "Prefectural Office" (都道府県庁) point data.
4. デジタル庁 アドレス・ベース・レジストリ (ABR)
   • Website: https://www.digital.go.jp/policies/base_registry_address
   • Format: CSV (Zipped)
   • Geometry Type: Point
   • Remarks:
     • Pinpoint Data for selected areas.

Notes

• For Bulk Geocoding, wherein addresses located in a field of a table are geocoded, please see this WIKI Entry.

• To create TRIGGERS that will geocode addresses automatically on an INSERT or UPDATE operation, please see this WIKI Entry.

• To do Spatial Analytics using pgGeocoder with GeoFuse, please see this WIKI Entry

• To view the FOSS4G Presentation for pgGeocoder, please see this WIKI Entry