{"id":7642,"date":"2023-10-11T17:49:18","date_gmt":"2023-10-11T15:49:18","guid":{"rendered":"https:\/\/www.geopostcodes.com\/en-GB\/?p=7642"},"modified":"2026-04-17T11:31:44","modified_gmt":"2026-04-17T09:31:44","slug":"geocoding-building-a-postcode-to-coordinates-converter","status":"publish","type":"post","link":"https:\/\/www.geopostcodes.com\/en-GB\/blog\/geocoding-building-a-postcode-to-coordinates-converter\/","title":{"rendered":"Geocoding: Building a Postcode to Coordinates Converter"},"content":{"rendered":"\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

<\/p>\n

Depending on the country, a postcode might be called a postcode, code postal, c\u00f3digo postal, PLZ, PIN code, \u90ae\u653f\u7f16\u7801, and the list undoubtedly goes on.<\/p>\n

<\/p>\n

<\/p>\n

Whatever you call it, a\u00a0postcode<\/em>\u00a0is a sequence of alphanumeric characters, sometimes formatted with spaces or hyphens. The primary use of postcodes is to facilitate the routing and sorting of letters and parcels, but they\u2019re also used for supply chain logistics, address cleaning, statistical research, sales, marketing, spatial analysis\u2026and again, the list continues.<\/p>\n

<\/p>\n

<\/p>\n

For each of those reasons, a business would need to identify the location of the postcodes. We\u2019ll walk you through a few use cases, discuss some of the challenges of converting postcodes to coordinates, and finally, how to build your own postcode to coordinates converter.<\/p>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

What\u2019s the Use of a Postcode to Coordinates Converter?<\/h2>\n

<\/p>\n

<\/p>\n

If you know the coordinates of a postcode, you can determine the price of delivery or service level terms. You can identify which areas can be reached in one business day or two. It can also be convenient information to relate offer and demand based on distance criteria, for example for jobs.<\/p>\n

<\/p>\n

<\/p>\n

Converting postcodes to coordinates is also critical for building maps and performing spatial analysis<\/a>, like looking for geographical patterns in your customer base. Are your customers mainly coming from specific areas or directions? How does that relate to transportation infrastructure, socio-demographic patterns, the presence (or absence) of competition, and so on?<\/p>\n

<\/p>\n

<\/p>\n

To identify the physical location of a postcode, you need a tool that assigns a coordinate to every (geographic) postcode; something that converts a postcode into a set of coordinates.<\/p>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

\n

<\/p>\n

Bear in mind that some postcodes are non-geographic (like the H0 H0 H0 for Santa Claus in Canada).<\/p>\n

<\/p>\n<\/blockquote>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

What Are the Challenges of Converting Postcodes to Coordinates?<\/h2>\n

<\/p>\n

<\/p>\n

Postcode granularity differs widely from country to country. It can directly identify a building, a part of a street, a town district, an entire town, or even a larger zone such as an administrative region. That means you\u2019ll have a lot of different types of data to process for an effective converter. The converter will have to deal with polygons, lines, and points, and convert them into a coordinate.<\/p>\n

<\/p>\n

<\/p>\n

The logical choice would be to use the \u201ccenter\u201d or the centroid of the data representing the postcode. But unfortunately, nothing is simple. There are cases where the geometric centroid of a polygon does not fall into that polygon:<\/p>\n

<\/p>\n

<\/p>\n

\"Centroid<\/figure>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

The postcode could be composed by several polygons:<\/p>\n

<\/p>\n

<\/p>\n

\"A<\/figure>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

Or only points:<\/p>\n

<\/p>\n

<\/p>\n

\"A<\/figure>\n

<\/p>\n

<\/p>\n

Clearly, you need a solution to ensure that the coordinate falls into the polygon or is one of the points.<\/p>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

So How Do You Build a Postcode to Coordinates Converter?<\/h2>\n

<\/p>\n

<\/p>\n

Software engineers know perfectly well what a coordinate is. It\u2019s a number that determines the position of a point in a space. Generally we use\u00a0x<\/em>,\u00a0y<\/em>, and sometimes\u00a0z<\/em>, but there are others. For example, the\u00a0Geographic Coordinate System (GCS)<\/a>.<\/p>\n

<\/p>\n

<\/p>\n

\u201cOh right,\u201d you say, \u201cI\u2019ve seen Google Maps, I know what latitude and longitude look like.\u201d<\/p>\n

<\/p>\n

<\/p>\n

Latitude and longitude are angles that refer to a point on an ellipsoid (an abstraction of the earth), but even they are not the only way to represent geographic coordinates.<\/p>\n

<\/p>\n

<\/p>\n

To know exactly what kind of reference is used for the coordinates, you have to know the\u00a0geodetic datum<\/a>. The one used by Google Maps and GPS is called the\u00a0World Geodetic System<\/a>\u00a0or WGS 84.<\/p>\n

<\/p>\n

<\/p>\n

GCSes are usually identified by their\u00a0EPSG<\/a>\u00a0reference. For WGS 84, it\u2019s EPSG:4326. Another commonly used reference is EPSG:3857, also called\u00a0Pseudo-Mercator<\/a>. It\u2019s projected onto a square and widely used on maps you see everywhere online.<\/p>\n

<\/p>\n

<\/p>\n

\n

<\/p>\n

Of course, there are many other reference systems that are specific to countries or regions; in this article, we\u2019ll show you how to transform data from one system to another.<\/p>\n

<\/p>\n<\/blockquote>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

Select Your Tools<\/h3>\n

<\/p>\n

<\/p>\n

It\u2019s time to define which tools you\u2019ll be using for your postcode to coordinates converter. You\u2019ll need to store data, process it, extract a pseudo-centroid for different collections of data (polygons or points), and visualize it.<\/p>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

Storing and Handling Data<\/h4>\n

<\/p>\n

<\/p>\n

As you\u2019re dealing with geographical coordinates, it seems obvious that you need a Geographical Information System (GIS) and a place to store data to work with. Why not a database?<\/p>\n

<\/p>\n

<\/p>\n

A perfect (and free) choice is PostgreSQL with the PostGIS extension. We won\u2019t cover the PostgreSQL installation here, but once you\u2019ve installed it, you only need to execute\u00a0CREATE EXTENSION postgis;<\/code>\u00a0in a psql prompt to set up PostGIS.<\/p>\n

<\/p>\n

<\/p>\n

\n

<\/p>\n

Note that other database systems propose similar solutions.<\/p>\n

<\/p>\n<\/blockquote>\n

<\/p>\n

<\/p>\n

\u00a0<\/div>\n

<\/p>\n

<\/p>\n

Computing Pseudo-centroids<\/h4>\n

<\/p>\n

<\/p>\n

A common way to find the centroid and to constrain it to be inside the geometry is to use the PostGIS function\u00a0ST_PointOnSurface<\/a>. It\u2019s not the only algorithm available and not always the best choice, but it\u2019s a good default.<\/p>\n

<\/p>\n

<\/p>\n

Let\u2019s try it with the previous examples we illustrated:<\/p>\n

<\/p>\n

<\/p>\n

SELECT ST_PointOnSurface(g), g\nFROM ST_GeomFromText('POLYGON((0 0, 0 1, 0.2 1, 0.2 0.2, 0.8 0.2, 0.8 1, 1 1, 1 0, 0 0))') AS g;\n<\/pre>\n
<\/p>\n
<\/p>\n
\"Coordinate<\/figure>\n
<\/p>\n
<\/p>\n
SELECT ST_PointOnSurface(g) , g\nFROM ST_GeomFromText('MULTIPOLYGON(((0 0, 0 1, 1 1, 1 0, 0 0),(2 0, 2 1, 3 1, 3 0, 2 0)))') AS g;\n<\/pre>\n
<\/p>\n
<\/p>\n
\"Coordinate<\/figure>\n
<\/p>\n
<\/p>\n
SELECT ST_PointOnSurface(g), g\nFROM ST_GeomFromText('MULTIPOINT((0 0), (0 0.21), (0.5 0.6), (0.4 0.3))') AS g;\n<\/pre>\n
<\/p>\n
<\/p>\n
\"Coordinate<\/figure>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Visualizing Progress<\/h4>\n
<\/p>\n
<\/p>\n
This tutorial uses\u00a0QGIS<\/a>\u00a0to load and view data, but you can also use geometry viewers of your favorite Postgres tool, like\u00a0PgAdmin<\/a>\u00a0or\u00a0DBeaver<\/a>.<\/p>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Take Your Tools to the Field<\/h3>\n
<\/p>\n
<\/p>\n
Now it\u2019s time to apply your tools to real-world examples!<\/p>\n
<\/p>\n
<\/p>\n
\n
<\/p>\n
A big challenge here is to find accurate and reliable sources. This tutorial only uses open data; post companies in some countries still sell them under license.<\/p>\n
<\/p>\n<\/blockquote>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Polygons<\/h4>\n
<\/p>\n
<\/p>\n
For the first use case, let\u2019s use\u00a0Postcode Tabulation Areas from the census bureau of the United States<\/a>. Click\u00a0Download<\/strong>\u00a0for the Shapefile Postcode File, unzip it, and drop the\u00a0.shp<\/code>\u00a0file into QGIS.<\/p>\n
<\/p>\n
<\/p>\n
It should look like this (I\u2019ve added an OpenStreetmap background to have some context):<\/p>\n
<\/p>\n
<\/p>\n
\"Viewing<\/figure>\n
<\/p>\n
<\/p>\n
To insert the polygons into a Postgres database, use command line tools such as\u00a0shp2pgsql<\/a>,\u00a0ogr2ogr<\/a>, or QGIS (via the menu\u00a0Processing > Toolbox > Database > Export to PostgreSQL<\/strong>).<\/p>\n
<\/p>\n
<\/p>\n
\n
<\/p>\n
As mentioned, this tutorial uses the QGIS processing, but in a production environment, you\u2019ll probably use command line tools that can be scripted.<\/p>\n
<\/p>\n<\/blockquote>\n
<\/p>\n
<\/p>\n
Name your table (this tutorial uses\u00a0us_zcta<\/code>\u00a0for the table name). Next, you have to add a geometry column to store the centroid of each postcode. It\u2019s constrained to store only point geometry with the WGS84 coordinate system (4326).<\/p>\n
<\/p>\n
<\/p>\n
\n
<\/p>\n
You can find more information on the\u00a0coordinate systems in PostGIS here<\/a>.<\/p>\n
<\/p>\n<\/blockquote>\n
<\/p>\n
<\/p>\n
ALTER TABLE us_zcta ADD COLUMN pos geometry(point, 4326);\n<\/pre>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
To compute the centroid, perform a transformation from the source spatial reference system; in this case, transform\u00a04269<\/code>\u00a0into the column we added, which is\u00a04326<\/code>.<\/p>\n
<\/p>\n
<\/p>\n
UPDATE us_zcta\nSET pos = ST_Transform(ST_PointOnSurface(geom), 4326);\n<\/pre>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Add the column\u00a0us_zcta.pos<\/code>\u00a0to QGIS to see the following result:<\/p>\n
<\/p>\n
<\/p>\n
\"Viewing<\/figure>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
To extract the latitude \/ longitude, perform the following select:<\/p>\n
<\/p>\n
<\/p>\n
SELECT zcta5ce10 AS postcode, ST_Y(pos) AS latitude, ST_X(pos) AS longitude \nFROM us_zcta;\n<\/pre>\n
<\/p>\n
<\/p>\n
\n
<\/p>\n
Note that the file already contained a latitude\/longitude column.<\/p>\n
<\/p>\n<\/blockquote>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Points<\/h4>\n
<\/p>\n
<\/p>\n
For the next example, let\u2019s go to Austria. Addresses with postcodes are available from the\u00a0Bundesministerium f\u00fcr Arbeit und Wirtschaft.\u00a0Download the postcode file directly<\/a>\u00a0and unzip it.<\/p>\n
<\/p>\n
<\/p>\n
The postcode contains a series of\u00a0.csv<\/code>\u00a0files. There are no geometric formats, so you\u2019ll have to load flat data first and then create the corresponding geometry.<\/p>\n
<\/p>\n
<\/p>\n
First, create a table:<\/p>\n
<\/p>\n
<\/p>\n
CREATE TABLE at_plz (\n    adrcd varchar,\n    gkz varchar,\n    okz varchar,\n    plz varchar,\n    skz varchar,\n    zaehlsprengel varchar,\n    hausnrtext varchar,\n    hausnrzahl1 varchar,\n    hausnrbuchstabe1 varchar,\n    hausnrverbindung1 varchar,\n    hausnrzahl2 varchar,\n    hausnrbuchstabe2 varchar,\n    hausnrbereich varchar,\n    hnr_adr_zusammen varchar,\n    gnradresse varchar,\n    hofname varchar,\n    rw double precision,\n    hw double precision,\n    epsg integer,\n    quelladresse varchar,\n    bestimmungsart varchar\n);\n<\/pre>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Open a\u00a0psql<\/code>\u00a0session and load the\u00a0ADRESSE.csv<\/code>:<\/p>\n
<\/p>\n
<\/p>\n
\\copy at_plz from 'ADRESSE.csv' with delimiter ';' quote '\"' CSV HEADER;\n<\/pre>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Add a geometry column and create the point from the\u00a0.csv<\/code>. There is a little subtlety here\u2014the file uses three different coordinate systems, as indicated in the\u00a0epsg<\/code>\u00a0column.<\/p>\n
<\/p>\n
<\/p>\n
ALTER TABLE at_plz ADD COLUMN geom geometry(point, 4326);\nUPDATE at_plz SET geom = ST_Transform(st_SetSRID(st_makepoint(rw, hw), epsg), 4326);\n<\/pre>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
And here is the table in QGIS. The points are automatically colored according to the postcode, to give you an idea of the zones they cover.<\/p>\n
<\/p>\n
<\/p>\n
\"Color-coded<\/figure>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Now, you have to compute the centroids of each postcode. As there are many rows with the same postcode, group them beforehand and save the results in a new table:<\/p>\n
<\/p>\n
<\/p>\n
CREATE TABLE at_plz_agg AS \nSELECT plz, ST_PointOnSurface(ST_Collect(geom))::geometry(point, 4326) pos\nFROM at_plz\nGROUP BY plz;\n<\/pre>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
View again in QGIS, note the red dots, and you\u2019re done!<\/p>\n
<\/p>\n
<\/p>\n
\"Data<\/figure>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
No Coordinates<\/h4>\n
<\/p>\n
<\/p>\n
Sometimes\u2014or rather, more often than sometimes\u2014a post company publishes postcodes without coordinates. In that situation, you have to cross the data with other sources to extrapolate the coordinates. Usually, countries have geographic or statistical institutes that maintain such references, or you can also use\u00a0OpenStreetMap<\/a>\u00a0or\u00a0GeoNames<\/a>\u00a0in some circumstances.<\/p>\n
<\/p>\n
<\/p>\n
Remember, when using multiple sources, have caution! You\u2019ll face matching problems, erroneous data,\u00a0normalization issues<\/a>, multiple languages and scripts, and other complications. Addressing these issues may require many steps, such as exploring and cleaning the data and removing outliers, just to name the first few.<\/p>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Conclusion<\/h2>\n
<\/p>\n
<\/p>\n
While you\u2019ve learned how to convert multiple object types into coordinates, specifically postcodes and their data sources and particularities, your converter is still a bit unfinished.<\/p>\n
<\/p>\n
<\/p>\n
To create an exhaustive worldwide dataset, you\u2019d need to centralize multiple countries in one table and keep the data up to date. There are about 180 countries using postcodes; some are updated yearly, some monthly, some weekly\u2026<\/p>\n
<\/p>\n
<\/p>\n
But don\u2019t be discouraged! You can work with a provider to handle those details for you. Like GeoPostcodes<\/a>.<\/p>\n
<\/p>\n
<\/p>\n
We maintain a worldwide list of postcodes and related data, from coordinates to administrative divisions, including geocodes and time zones. We\u2019ve updated our data hundreds of times and built advanced processing pipelines to deal with the particularities of each country. For example, 99.6 percent of our postal codes have accurate coordinates. The remaining 0.4 percent inherit coordinates from the administrative divisions they belong to. These are the problems we enjoy solving, and we\u2019d be happy to save you the trouble! Reach out to us<\/a> for further information, or you can directly explore our data<\/a> from our website.<\/p>\n
<\/p>\n
<\/p>\n
\u00a0<\/div>\n
<\/p>\n
<\/p>\n
Additional Resources<\/h3>\n
<\/p>\n
<\/p>\n