Updates docs for postal code
This commit is contained in:
Carla Iriberri
@@ -2,11 +2,14 @@ Postal code geocoder
|
||||
===============
|
||||
|
||||
This section is divided in geocoding postal codes as points or as polygons. Each option has its own sources and its own functions which are described below.
|
||||
|
||||
# Postal code geocoder: Polygons
|
||||
## Function
|
||||
|
||||
By following the next steps a table is populated with zipcodes from Australia, Canada, USA and France (identified by iso3) related with their spatial location in terms of polygons.
|
||||
|
||||
## Usage example
|
||||
|
||||
## Creation steps
|
||||
|
||||
1. Import the four files attached in the section "Datasources" for Australia (`doc` table), Canada (`gfsa000a11a_e` table), USA (`tl_2013_us_zcta510` table) and France (`codes_postaux` table).
|
||||
@@ -18,15 +21,15 @@ By following the next steps a table is populated with zipcodes from Australia, C
|
||||
#### Table structure
|
||||
````
|
||||
Table "public.postal_code_polygons"
|
||||
Column | Type | Modifiers | Storage | Stats target | Description
|
||||
Column | Type | Modifiers | Storage | Stats target | Description
|
||||
----------------------+--------------------------+-----------------------------------------------------------------------+----------+--------------+-------------
|
||||
cartodb_id | integer | not null default nextval('untitled_table_2_cartodb_id_seq'::regclass) | plain | |
|
||||
postal_code | text | | extended | |
|
||||
adm0_a3 | text | | extended | |
|
||||
created_at | timestamp with time zone | not null default now() | plain | |
|
||||
updated_at | timestamp with time zone | not null default now() | plain | |
|
||||
the_geom | geometry(Geometry,4326) | | main | |
|
||||
the_geom_webmercator | geometry(Geometry,3857) | | main | |
|
||||
cartodb_id | integer | not null default nextval('untitled_table_2_cartodb_id_seq'::regclass) | plain | |
|
||||
postal_code | text | | extended | |
|
||||
adm0_a3 | text | | extended | |
|
||||
created_at | timestamp with time zone | not null default now() | plain | |
|
||||
updated_at | timestamp with time zone | not null default now() | plain | |
|
||||
the_geom | geometry(Geometry,4326) | | main | |
|
||||
the_geom_webmercator | geometry(Geometry,3857) | | main | |
|
||||
|
||||
````
|
||||
|
||||
@@ -50,7 +53,7 @@ Indexes:
|
||||
public | geocode_postalcode_polygons | SETOF geocode_namedplace_v1 | code text[] | normal
|
||||
public | geocode_postalcode_polygons | SETOF geocode_namedplace_country_v1 | code text[], inputcountries text[] | normal
|
||||
public | geocode_postalcode_polygons | SETOF geocode_namedplace_v1 | code text[], inputcountry text | normal
|
||||
`````
|
||||
````
|
||||
|
||||
### test_geocode_postalcode_polygons
|
||||
````
|
||||
@@ -58,7 +61,7 @@ Indexes:
|
||||
--------+----------------------------------+-------------------------------------+------------------------------------+--------
|
||||
public | test_geocode_postalcode_polygons | SETOF geocode_namedplace_country_v1 | code text[], inputcountries text[] | normal
|
||||
public | test_geocode_postalcode_polygons | SETOF geocode_namedplace_v1 | code text[], inputcountry text | normal
|
||||
`````
|
||||
````
|
||||
|
||||
## Response data types
|
||||
* geocode_namedplace_country_v1:
|
||||
@@ -71,12 +74,20 @@ Indexes:
|
||||
## Data Sources
|
||||
|
||||
* Australian polygons - http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/2033.0.55.0012011?OpenDocument - Download the KMZ for *Postal Area IRSD, SEIFA 2011*. Unzip and upload the kmz
|
||||
- Coverage: AUS
|
||||
- Geometry type: polygon
|
||||
|
||||
* Canadian polygons - http://www12.statcan.gc.ca/census-recensement/2011/geo/bound-limit/bound-limit-2011-eng.cfm - Download ESRI Shp, Forward Sortation Areas, Digital Boundary
|
||||
* Canadian polygons - http://www12.statcan.gc.ca/census-recensement/2011/geo/bound-limit/bound-limit-2011-eng.cfm - Download ESRI Shp, Forward Sortation Areas, Digital Boundary
|
||||
- Coverage: CAN
|
||||
- Geometry type: polygon
|
||||
|
||||
* USA polygons - http://www2.census.gov/geo/tiger/TIGER2013/ZCTA5/tl_2013_us_zcta510.zip
|
||||
- Coverage: USA
|
||||
- Geometry type: polygon
|
||||
|
||||
* French polygons - http://www.data.gouv.fr/dataset/fond-de-carte-des-codes-postaux
|
||||
- Coverage: FRA
|
||||
- Geometry type: polygon
|
||||
|
||||
# Postal code geocoder: Points
|
||||
|
||||
@@ -93,6 +104,8 @@ IT, LU, SK, LI, PR, IM, NO, PT, PL, FI, JP, CA, DE, HU, PH, SE, VA, YT, MK, FR,
|
||||
MH, RO, FO, GF, AD, HR, DZ, GT, AU, AS, BE, AT
|
||||
````
|
||||
|
||||
## Usage example
|
||||
|
||||
## Creation steps
|
||||
|
||||
1. Download the allCountries.zip file from [GeoNames](www.geonames.org). Import and rename the table as tmp_zipcode_points. You can follow the manual process explained below instead.
|
||||
@@ -121,15 +134,15 @@ Import the file ignoring step 2.
|
||||
#### Table structure
|
||||
````
|
||||
Table "public.postal_code_points"
|
||||
Column | Type | Modifiers | Storage | Stats target | Description
|
||||
Column | Type | Modifiers | Storage | Stats target | Description
|
||||
----------------------+--------------------------+------------------------------------------------------------------------+----------+--------------+-------------
|
||||
cartodb_id | integer | not null default nextval('untitled_table_2_cartodb_id_seq2'::regclass) | plain | |
|
||||
adm0_a3 | text | | extended | |
|
||||
postal_code | text | | extended | |
|
||||
created_at | timestamp with time zone | not null default now() | plain | |
|
||||
updated_at | timestamp with time zone | not null default now() | plain | |
|
||||
the_geom | geometry(Geometry,4326) | | main | |
|
||||
the_geom_webmercator | geometry(Geometry,3857) | | main | |
|
||||
cartodb_id | integer | not null default nextval('untitled_table_2_cartodb_id_seq2'::regclass) | plain | |
|
||||
adm0_a3 | text | | extended | |
|
||||
postal_code | text | | extended | |
|
||||
created_at | timestamp with time zone | not null default now() | plain | |
|
||||
updated_at | timestamp with time zone | not null default now() | plain | |
|
||||
the_geom | geometry(Geometry,4326) | | main | |
|
||||
the_geom_webmercator | geometry(Geometry,3857) | | main | |
|
||||
````
|
||||
#### Current indexes
|
||||
````
|
||||
@@ -152,6 +165,8 @@ Import the file ignoring step 2.
|
||||
## Data Sources
|
||||
|
||||
* All countries points [GeoNames](www.geonames.org) - http://download.geonames.org/export/zip/allCountries.zip
|
||||
- Coverage: See details at function description
|
||||
- Geometry type: point
|
||||
|
||||
# Geocoder coverage map
|
||||
|
||||
@@ -160,22 +175,24 @@ Import the file ignoring step 2.
|
||||
|
||||
# Known deficiencies of the service
|
||||
* For the USA polygon zipcode service, Zipcode Tabulation Areas (ZCTA) are being used which [don't correspond to actual zipcode regions](http://web.archive.org/web/20050209030255/http://www.manifold.net/cases/zip_codes/zip_codes.html).
|
||||
* Regarding the point geocoder service, being offered from GeoNames data: we've detected that the accuracy for a big section of zipcodes is not as good as intended, as GeoNames interpolates zipcode-populated place information. As an example, in the case of Madrid, Spain, all the zipcodes belonging to the city are geocoded in the centroid of the city itself.
|
||||
* Regarding the point geocoder service, being offered from GeoNames data: we've detected that the accuracy for a big section of zipcodes is not as good as intended, as GeoNames interpolates zipcode-populated place information. As an example, in the case of Madrid, Spain, all the zipcodes belonging to the city are geocoded in the centroid of the city itself.
|
||||
This issue can be spotted easily by comparing interesecting zipcode points:
|
||||
`SELECT the_geom, the_geom_webmercator, COUNT(*), adm0_a3 FROM postal_code_points GROUP BY the_geom, the_geom_webmercator, adm0_a3 HAVING COUNT(*) > 1 order by count(*)`
|
||||
|
||||
In this case, we conclude that most affected countries are Portugal, Mexico, Spain, Netherlands, Czech Republic or Slovakia, meanwhile Brazil doesn't show intersecting values.
|
||||
|
||||
|
||||
The visual result of intersecting zipcodes is demonstrated in the following figure:
|
||||
|
||||
|
||||
# Historic:
|
||||
* [19/10/2015]:
|
||||
* Updates readme with usage examples and setup scripts
|
||||
* [08/10/2015]:
|
||||
* Adds response data types
|
||||
* Adds response data types
|
||||
* [15/07/2015]:
|
||||
* Adds basic tests for postal codes polygons
|
||||
* [02/07/2015]:
|
||||
* Adds known deficiencies and coverage map
|
||||
* Adds known deficiencies and coverage map
|
||||
* [24/06/2015]:
|
||||
* Updated readme.md: added information about tables, function, indexes and the known issues section.
|
||||
* Updated readme.md: added information about tables, function, indexes and the known issues section.
|
||||
* Review and [upload functions](https://github.com/CartoDB/data-services/pull/152)
|
||||
|
||||
5
geocoder/setup/postal-codes/response-types.sql
Normal file
5
geocoder/setup/postal-codes/response-types.sql
Normal file
@@ -0,0 +1,5 @@
|
||||
-- Response types for postal codes geocoder
|
||||
CREATE TYPE geocode_namedplace_v1 AS (q TEXT, geom GEOMETRY, success BOOLEAN);
|
||||
CREATE TYPE geocode_place_country_iso_v1 AS (iso3 TEXT, c TEXT, q TEXT, geom GEOMETRY, success BOOLEAN);
|
||||
CREATE TYPE geocode_namedplace_country_v1 AS (c TEXT, q TEXT, geom GEOMETRY, success BOOLEAN);
|
||||
CREATE TYPE geocode_postalint_country_v1 AS (c TEXT, q INTEGER, geom GEOMETRY, success BOOLEAN);
|
||||
@@ -0,0 +1 @@
|
||||
-- Triggers for postal codes geocoder
|
||||
|
||||
Reference in New Issue
Block a user