Postal Code Musings « Liliendahl on Data Quality


12 January 2013 11:10

Postal CodeMusings

When working with master data management and data quality including data matching one of the most frequent pieces of information you work with is a postal code.

Wikipedia has a good article about postal code.

Some of the data quality issues related to the datum postal code are:

Metadata

Over the world different words are used for a postal code:

· ZIP code, the United States implementation of a postal code, is often used synonymously for a postal code in many databases and user interfaces. This is not seriously wrong, but not right either.

· In India a postal code (in English) is called a PIN Code (Postal Index Number). This could definitely trick me.

Format

There are basically two different formats of postal codes around:

· Numeric postal codes are the most common ones. The number of digits does however differ between countries. And there may be some additional considerations:

o For example the 9 digit United States ZIP code is split into the original 5 digits and the additional 4 digits implemented later.

o Postal codes may begin with 0 which may create formatting errors when treated as numeric.

· Some countries, for example the United Kingdom, the Netherlands, Canada and Argentina, have alphanumeric postal codes.

Embedded Information

Numeric postal codes usually forms some kind of hierarchy in which you can guess the geographical position within the country and make ranges representing smaller or larger geographical areas. But you never know.

This also goes for Dutch (you know, t he ones in the Netherlands) postal codes as the first 4 characters are numeric.

The UK postal codes usually start with a mnemonic of the main city in the area, except in a lot of cases.

Precision

Some postal codessystems have postal codescovering larger areas with many streets and some postal code systems are very granular where each street, or part of a street, has a distinct postal code.

The UK postal code system is very granular which have paved the way for using rapid addressing as told in a recent article on the UK Database Marketing Magazine .

Coverage

Utilizing rapid addressing requires that reference data for postal codes practically covers every spot in the c ountry and updates are available on a near real time basis.

Some countries have postal codes systems not covering every corner and some countries haven &#146 t a postal code system at all.

Uniqueness

The main reason for implementing postal code systems is that a town or city name in many cases isn &#146 t unique within a country.

But that doesn &#146 t mean that uniqueness works the other way as well. A postal code may in many countries cover several town names. France is an example.

Consistency

While we basically have granul ar and not so granular postal code systems we of course also have hybrids.

In Denmark for example there is a granular system in the capital Copenhagen with a postal code for each street, named by the street, and a system in the rest of country with a posta l code for an area named by the suburban or town.

Fit for purpose

A postal code is a hierarchical element in a postal address. We basically have two forms of postal addresses:

· A geographical address where the postal address including the postal code points to place you also can visit and meet the people receiving the things sent to there

· A post-office box which may have more or less geographical connection to where the people receiving the things sent to there are

Penetration of post-office boxes differs around the world. In Namibia it is mandatory. In Sweden most companies have a post-office box address.

Trying to compare data with these different concepts is like comparing apples and oranges, which often goes bananas .

5 Responses to Postal CodeMusings

1. Graham Rhind says: 10th January 2013 at 16:30 Nice one, Henrik.Just a note that “post-office box” address is generically better referred to as a “mailing” address, as non-geographic addresses can include postbags, rural routes, postal drawers and a whole host of other forms. I would also include a third address form: a large-user address,. where the company name alone, possibly along with a code or place name, is sufficient for delivery. In some countries a large company can have three postal codes &#150 one for the company, one for its mailing address and one for its geographical address. Reply

o Henrik Liliendahl Sørensen says: 10th January 2013 at 16:52 Thanks Graham. As always good in depth knowledge on international addressing from your side. Reply

2. Gary Palmer says: 11th January 2013 at 10:07 I stand ready to be corrected by Graham, who is wiser than I on these things, but I believe that the UK postal codes usually start not with a mnemonic of merely the main city in the area, but a code representing the main or hub postal sorting office for that area &#150 thus, for example, all post for Postcodes beginning TN will pass through the sorting office at Tonbridge Reply

3. Graham Rhind says: 11th January 2013 at 11:00 Quite right Gary Reply

4. Henrik Liliendahl Sørensen says: 11th January 2013 at 11:28 There you see. I think it will be hard to rationalize the sorting system if this connection has to be maintained.I have found that the British cherish their postcodes in level with the London Taxi and Sunday Roast

Inserted from < liliendahl.com/2013/01/10/postal-code-musings/ >



Practical International Data Management Online.  A free resource from GRC Data Intelligence. For comments, questions or feedback: pidm@grcdi.nl