Google Chrome Translate trying to translate maps

It thinks this is Norwegian

It thinks this is Norwegian

The original

The original

The minimal location entity set

A minimal location entity set is an alphanumeric description of a location which maps to a single point (Latitude, Longitude) without ambiguity, and is the shortest such description.
For example:
La Fàbrica del Sol, Passeig de Joan Salvat Papasseit, 1, 08003 Barcelona, Spain


While many geocoding systems attempt to return as much data as possible (such as alternative road names, neighborhood name, timezones, geohashes, Ordnance Survey gridrefs, calling codes, what3words, sunrise/sunset, and more) all this seems unnecessary since you may get this additional information of you perform a reverse geocoding lookup on the point.

Returning so much redundant information on an address lookup seems like a waste of space to me.

Let’s build a two step geocoding system, one that a) returns the most minimal info on the first request and b) returns all the rest of location details on the second.

Geocoder vs CanadaPost Lawsuit Updates
Lawsuit Update

CanadaPost’s lawsuit, now in its 4rth year, is ongoing and it looks like it is finally getting a court date soon (they have been quiet for a while, probably wishing to keep this under wraps until after the Federal Election.)

Either way, we are still here, and we are still providing a free database of postal addresses and postal codes that is bigger and better than ever.

The main database has grown considerably in the last 4 years, further proof that crowdsourcing works! As of the last update on 2015-09-30, 12613 new postal codes were added, with the total now approaching 1 million.

You may download all these data for free at (under the Creative Commons Attribution 2.5 Canada License.)

Will keep you posted as to the latest developments from the federal court. All the best and thank you for your support.

Ervin Ruci

@Yapc::EU 2015 explaining fuzzy geocoding…

@Yapc::EU 2015 explaining fuzzy geocoding with

K2C Polygons

I just got around to comparing the FSA Polygons that I have crowdsourced on with the ones that were released a few years back by Statistics Canada after they talked Canada Post into it.

The results are surprising. The “official” polygons for the FSA’s seem to leave out quite a bit of the area they actually cover. I know this for a fact, because K2C 1N5 is my postal code. The crowdsourced FSA’s, (available for free download here) are the best approximations for FSA polygons.

See for yourself:


I gave a talk at State of the Map conference on crowdsourcing. (The slides are posted on my twitter feed , also available on Vimeo)

One observation I took home from the conference is that the state of public data around the word is similar to that in Canada, in the sense that governments and their affiliate entities hold on to the data for as long a possible, despite the fact that doing so, adversely affects the state of their economies and goes against the public good. (According to OSM France “Bano” project, a country loses up to 0.5% of its GDP due to lack of publicly available addressing data. Source)

Crowdsourcing the data is not an optimal solution, in the face of the lack a data feed from its authoritative source, because it results in datasets that contain errors. Still, this seems the only way to open up the data in my view, when the decision makers are convinced that keeping the data closed is better for their budgets (an interesting figure of 0.5 billion pounds was thrown around as the value of a closed post code list by the Royal Mail CEO Moya Greene, in her arguments for keeping the dataset closed. She also happened to be Canadapost’s CEO at the time they started their legal efforts aimed at enforcing CP’s alleged intellectual rights over Canadian postal codes.)

In France on the other hand they don’t have such problems, as they have not yet made the effort to create a post code system like the Canadian or the British ones, hence it is hard to make the half billion dollar argument there, still that does not mean that whatever system they have is open. People from the “Bano” project had to lobby hard to get the list of up to 1000 postal codes created by the French postal service open to the public, and when they actually did it was full of errors. Not only that but the French postal service has 4 different street address datasets (one for regular mail, one for advertising mail,  one for parcels and another one for a purpose I can’t remember now.) All 4 have quality issues, and the 4 different departments that created them do not talk nor cooperate with each other to improve their respective datasets. Funny stories of government inefficiency at the public’s expense. The Economist also wrote a piece on this topic a month ago.

In closing, public data all around the world is at various stages of unavailability because certain people of influence are convinced they are worth a lot of money. Nobody has yet shown how much money they are actually making from licensing this data. I doubt it is half a billion. Or 0.5% of the GDP.

I am certain it is more akin to a hidden tax we all have to pay.


T0E P0R8L3 C0W

I wrote a poem then translated it into postal code. Enjoy!

1H1A0A N3V3R1 S3E P0R8L3 R0C0W0

A2N1N0 N0T T1N0K0 0N3 E6X1S1

B0T 1F1 T0E Y0U B3T3R3 L0K0T0

N0T T0 M1S T0K F0R P0R8L3 S3E3E3


With apologies to:

The Purple Cow. 

I have never seen a purple cow

and i do not think that one exist

but if it does you’d better look out

not to mistake it for a purple sheep