Geocoding naming formats and challenges

From OpenStreetMap Wiki
Jump to: navigation, search

One of the hard things about geocoding is parsing addresses into something intelligible. This list was started to help people who are building geocoders and address parsers. The list was started by user:nickb.

Please add any examples you have and help structure the list. To make a good address parser its useful to know the different ways people search for addresses in different countries.

Countries

Aliases are often used for countries:

  • Brazil (Brasil), br, bra
  • Canada, ca, can
  • Chile, cl
  • UK, United Kingdom, Great Britain, Britain, British Isles
  • United States, us, usa
  • South Africa, Republic of South Africa, RSA, za

Austria (Österreich)

Addresses

Addresses without a street name (mostly referring to a village or hamlet name):

States (Bundesländer)

  • Wien
  • Tirol

(not commonly used in Addressing)

Cities/Villages

  • Wien
  • Hall
  • Hall in Tirol
  • Hall i.T.
  • Sankt Jakob in Defereggen
  • St. Jakob in Defereggen
  • St. Jakob i. Def.
  • St. Jakob

Brazil (Brasil)

Addresses

  • Rua Projetada, RJ
  • Rua Projetada, Rio de Janeiro
  • Rua Projetada, 345, Rio de Janeiro
  • Rua Projetada, 345, Rio de Janeiro - RJ
  • Rua Projetada, 345, Rio de Janeiro - RJ, 20771-004
  • Rua Projetada, Rio de Janeiro - RJ, Brasil‎
  • Rua Projetada, Rio de Janeiro - RJ, 20771-004
  • Rua Projetada, Rio de Janeiro - RJ, Brasil‎
  • Rua Projetada, 345, Rio de Janeiro - RJ, Brasil‎
  • Rua Projetada, 345 apto 504, Rio de Janeiro - RJ, 20771-004, Brasil‎
  • Rua Projetada 345 apto 504 - Rio de Janeiro - RJ, 20771-004, Brasil‎

States (Estados)

  • Rio de Janeiro, RJ

Cities (Cidades)

  • Rio de Janeiro, RJ

Canada

Addresses

  • 1011 University Drive, Saskatoon
  • 1011 University Drive, SK
  • 1011 University Drive, Saskatoon, SK
  • 1011 University Drive, Saskatoon, Canada
  • 1011 University Drive, Saskatoon, SK, Canada
  • 1011 University Drive, Saskatchewan
  • 1011 University Drive, Saskatchewan, Canada
  • 1011 University Drive, Saskatchewan, CA
  • 1011 University Drive, Saskatoon, Saskatchewan
  • 1011 University Drive, Saskatoon, Saskatchewan, CA
  • 1011 University Drive, Saskatoon, SK S7N 0K4, CA
  • 1011 University Drive, Saskatoon, SK S7N 0K4, Canada

States (Provinces)

  • Saskatchewan, SK

(not commonly used in Addressing)

Cities

  • Saskatoon

Chile

Addresses

  • Cirujano Guzman 24, Departamento 102, Providencia, Santiago, Chile
  • Cirujano Guzman 24, Departamento 102, Providencia, Santiago
  • Cirujano Guzman 24, Departamento 102, Providencia
  • Cirujano Guzman 24, Depto 102, Providencia, Santiago, Chile
  • Cirujano Guzman 24, Depto. 102, Providencia, Santiago, Chile

Tipically (street name) (house number), (apartment) (apartment number), (municipality, a name that in most of the cities outside of Santiago is the same as the city name and in those cases you usually change it for the suburb name), (city), (country); in that order and most of the times without the Country (sometimes without the City, with well known municipalities inside Santiago).

Regions (Regiones)

(not commonly used in Addressing, mostly to identify towns with a name that repeats many times all over the country. Should be between the city name and the country)

  • Región Metropolitana de Santiago
  • XIII Región Metropolitana de Santiago
  • Región Metropolitana
  • RM


  • Región Aysén del General Carlos Ibáñez del Campo
  • XI Región Aysén del General Carlos Ibáñez del Campo
  • Región de Aysén
  • XI Región

Cities

Sometimes the city name is shortened or simply cuted in the middle because we think there is no other city with a similar name or just because we are lazy

  • Stgo
  • Stgo.

(For Santiago)

  • Viña

(For Viña del Mar)

  • San Pedro

(For San Pedro de Atacama, San Pedro de La Paz, San Pedro de Alcántara, Bahia San Pedro, etc. As you can see a huge mistake since there are just to many towns with the same name all over the country)

Colombia

Addresses

  • Calle 54 No 27-34
  • Clle. 54 # 27-34
  • CL 54 27 34


  • Carrera 23 No 35-40
  • Cra. 23 # 35-40
  • KR 23 35 40


Typically some other elements as floor, apartment number, neighborhood or building name are appended because the addressee name is usually not enough:

  • Cra 26 D # 74 B - 76 Apto 34
  • Kra 26D #74B-76 Apartamento 34
  • Carrera 26D 74B-76, Apt 34
  • Cra 26D # 74B - 76, Edificio Gasy, Apto 34, Barrio El Silencio
  • Cra 26D # 74B - 76, Edificio Gasy, Apto 34
  • Cra 26D # 74B - 76, Apt 34, Edif. Gasy


Grid System

The abstract idea

Addresses in Colombia intend to follow a Cartesian scheme (trying to use, where possible, only the 1st quadrant). Let's take an imaginary perfect square city. In this city, streets parallel to the NS axis (y axis) are called carreras and streets orthogonal to carreras (thus parallel to x axis) are called calles (Spanish for street). Both calles and carreras are equally spaced and increasingly numbered starting at 1 (calles from South to North and carreras from East to West). This way, we can easily locate a crossroads between a calle and a carrera or calculate distances between points. Now let's take an address in this city and see what we can infer from the information it gives us:

Carrera 5 # 9-25

Where 9-25 will be written on the house door.

It tells us the following:

  1. The house is located at 5th carrera
  2. The house is located North of the crossroads between the 5th carrera and the 9th calle
  3. The door is located at approx. 25m from that crossroads
  4. The house is located on the West side of the street (because 25 is an odd number)

The general case would be then:

[calle or carrera?] [number of that calle o carrera] # [number of the orthogonal street (the lesser street number from within the two orthogonal ones delimiting the segment]-[distance from crossroads in the direction of the numbering increment + street side (odd or even?)]

A real-life example of street numbering: http://www.openstreetmap.org/?lat=4.59621&lon=-74.07441&zoom=17&layers=B000

The actual result

Sometimes the streets are not perfect straight or solid lines. They don't always go from S to N or from W to E. Nonetheless, it's usually just a matter of turning over the plane, having several segments with different street numbers o several streets with the same street number. That said, and having the abstract idea in mind, we have to introduce some modifications in order to adapt the scheme to the natural growth of urban zones.

Not completely orthogonal streets: in this case, the street is called transversal if it approximates to the carrera-alignment, or diagonal if it approximates to the calle-alignment. Calles are considered orthogonal to transversales and carreras to diagonales.

Minor/new streets between already-numbered ones: if there is no possible number between the incumbent streets, then the new street takes the smallest one plus an ordered letter (A,B,C,...,Z) with exception of N, S, E, O (Spanish for W[est]).

Areas outside the 1st quadrant: the Cartesian scheme is conserved, only that, instead of negative numbers, whether the southern/eastern streets acquire an S/E afterwards, or the northern/western streets acquire an N/O, depending on the city. The full words Norte (North), Sur (South), Este (East) or Oeste (West) are also frequently used.

States (Departamentos)

  • Atlántico
  • Bolívar

Cities

  • Barranquilla, Atlántico
  • Barranquilla

Czech Republic

Addresses

  • Novosady 555/11, 78401 Litovel
  • Lidické náměstí 3, 400 01 Ústí nad Labem
  • Za Brumlovkou 266/2, 140 22 Praha 4 – Michle
  • Za Kaňkou 232, Příbram III. 261 01

Germany

Addresses

  • Jungfernstieg 20, Hamburg
  • Jungfernstieg 20, 20354 Hamburg
  • Jungfernstieg 16-20, 20354 Hamburg
  • Jungfernstieg 20, 20354 Hamburg, Deutschland
  • Jungfernstieg 20, 20354 Hamburg, Germany
  • Jungfernstieg 20, 2000 Hamburg 50 (obsolete, pre 5-digit ZIP codes)

States (Bundeslaender)

  • Bayern

Cities

  • Augsburg
  • Augsburg, Bayern


India

Addresses

Postal codes are standard there are Nine Pin zones to cover all the states and Military postal service you can read more about it here Indian Postal system

Urban Areas

Name Occupation (Optional) Flat Number, Building Name Street Number, Street Name Locality or Neighbourhood CITY - Postal Code (PIN) State

Rural Areas

Name Street Number, Street Name VILLAGE NAME District Name Postal Code (PIN) State


Ireland (Republic)

Addresses

Peculiarities are the absence of any postal codes (closest equivalent is Dublin postal district) and the prominence of counties.

Dublin

Except in outlying parts of the county, addresses are of the form: Street, Suburb/District, Postal District:

  • 42 John Street, Blanchardstown, Dublin 15

(Note that neither "Dublin 15", "15" nor the common abbreviation "D15" can usefully be deemed a postcode)

In the city centre, there may be no suburb or district:

  • 12 O'Connell Street, Dublin 1

Outlying parts of the county, along with all of the modern administrative county of Dunleary Rathdown, do not have Postal Districts:

  • 3 Main Street, Garristown, County Dublin
  • 14 The Downs, Blackrock, County Dublin

Note that, although these addresses are in the modern counties of, respectively, Fingal and DL/Rathdown, "County Dublin" is still used in addressing. Traditional counties enjoy much more public awareness than the modern subdivisions

"County" is frequently abbreviated to "Co." and always precedes the county name in addresses.

Urban Areas

Street, Town, County:

  • 4 Clare Street, Nenagh, County Tipperary

Terraces or other non-obvious subdivisions of streets may exist:

  • 9 Grattan Terrace, High Street, Tuam, County Galway

Rural Areas

Many areas have neither road names nor house numbers. Instead, addresses are expressed as Recipient Name, Townland, Postal Town, County:

  • John Murphy, Nuttstown, Dunboyne, County Meath
  • Welding Supply Stores, Johnswood, Manorhamilton, County Leitrim

Townlands are an established subdivision of land and are on OSM typically represented (where they are known) as place=locality

House numbers may exist:

  • Mary Manning, The Pines, Nuttstown, Dunboyne, County Meath

Boundary Conditions

Sometimes an address in one county can have a postal town in a different one. This gives rise to confusing addresses:

3 Little Pace Avenue, Clonee, Dublin 15

(Clonee lies entirely in County Meath)

Italy

Addresses

Typically: (street type) (street name), (house number) (5-digits postal code) (town) (two-letter code of province)

  • street type can be Via, Piazza, etc. (see Name_finder:Abbreviations#Italiano_-_Italian for some examples)
  • house number is preceded by a comma (or traditionally by the letter "n" or "n.") and a space
  • postal code is always 5 digits even if it starts with zero (eg. 00100 for Rome)
  • two-letter code of province is the same used on license plates, sometimes enclosed in round parenthesis

The Netherlands

Addresses

Plain address example Postal codes are always four numeral digits and two upper case letters (1234AB), but a space often used between digits and letters (1234 AB). House number extensions will be attached to the number, mostly without a space. In some cities, numbers are used as apartment indicators, indicating the floor number. Roman numerals are also used.

  • Boerhaavelaan 60, Doetinchem
  • Boerhaavelaan 60, 7002HX Doetinchem
  • Boerhaavelaan 60, 7002 HX Doetinchem
  • Boerhaavelaan 60, 7002 HX, Doetinchem

With extension

  • Steenstraat 12a, Arnhem
  • Steenstraat 12a, 6828CK Arnhem
  • Steenstraat 12a, 6828 CK, Arnhem

With apartment indication

  • Kanaalstraat 197-2, Amsterdam
  • Kanaalstraat 197-2, 1054XG Amsterdam
  • Kanaalstraat 197-2, 1054 XG, Amsterdam

Roman numerals

  • Kanaalstraat 197-II, 1054 XG, Amsterdam

Combination of floor and apartment

  • Kanaalstraat 197-2a, 1054 XG, Amsterdam

With special case suffix for "upper floor"

  • Kanaalstraat 197 bis, 1054 XG, Amsterdam

Just a postal code and (complete) house number with a space in between can occur, as this in essence makes street and city redundant

  • 7002HX 60 / 7002 HX 60
  • 6828CK 12a / 6828 CK 12 a
  • 1054XG 197-2 / 1054 XG 197-2

Provincies (States)

Not commonly used in addressing

  • Friesland
  • Groningen
  • Drenthe ("Drente" can also be seen)
  • Overijssel
  • Gelderland
  • Noord-Holland
  • Zuid-Holland
  • Utrecht
  • Zeeland
  • Noord-Brabant
  • Limburg
  • Flevoland

Cities

Can have a couple of aliases

  • Doetinchem
  • Den Haag = 's-Gravenhage
  • Den Bosch = 's-Hertogenbosch
  • Utrecht

South Africa

Addresses

  • 68 Kenridge Avenue, Durbanville, Cape Town 7550, South Africa
  • 68 Kenridge Avenue, Kenridge, Cape Town 7550, South Africa
  • 68 Kenridge Avenue, Durbanville, 7550 South Africa
  • 68 Kenridge Avenue, Durbanville, 7550, Republic of South Africa
  • 68 Kenridge Avenue, Cape Town, ZA

States (Provinces)

  • Western Cape
  • Gauteng
  • Kwazulu Natal
  • Mpumalanga
  • Free State
  • Northern Cape
  • Eastern Cape
  • North West Province
  • Limpopo

Cities

  • Cape town
  • Cape Town, Western Cape
  • Cape Town, South Africa
  • Durban, Kwazulu Natal
  • Durban, KZN
  • Durban

Spain

Addresses

Tipically (street name), (street number), (block), (stairwell), (floor+door), (postal code), (city), (province); in that order. Most addresses don't have a block (to tell apart several buildings with the same street number) nor stairwell.

  • Calle de Nuestra Señora de la Luz, nº 48, bajo C, 28047 Madrid, Madrid
  • Calle de Nuestra Señora de la Luz, nº 48, bº C
  • Calle de Nuestra Señora de la Luz, nº 48, bloque 3, escalera derecha, segundo izquierda, 28047 Madrid
  • Calle de Nuestra Señora de la Luz, nº 48, bl. 3, esc. dcha, 2 izq.
  • C/ Nuestra Señora de la Luz, nº 48, bl. 3, esc. dcha, 2 izq.
  • C/ Ntra. Sra. de la Luz, 48, bl. 3, esc. dcha, 2 izq.
  • C/ Ntra. Sra. de la Luz, 48

UK

  • 1-2 Hatfields, London
  • 1-2 Hatfields, London, SE1
  • 1-2 Hatfields, London, SE1 9PG
  • SE1 9PG
  • 1-2 Hatfields, London
  • 1-2 Hatfields, London, UK, SE1
  • 1-2 Hatfields, London, england, SE1 9PG
  • Marlin House, Putney, London
  • Marlin House, SW15
  • Marlin House, Putney
  • Flat 16, Marlin House, Putney, London
  • Flat 16, Marlin House, SW15
  • Flat 16, Marlin House, Putney
  • 26 Rayner's Road, Putney, London, UK
  • 26 Rayner's Road, Putney, London
  • 26 Rayner's Road, Putney, SW15
  • 24 PA18 6BD
  • Gorse Hill, Hook Heath Road, Woking, Surrey, GU22 0QH
  • Hook Heath Road, Woking, Surrey, England
  • Gorse Hill, Surrey, England
  • London
  • London, UK
  • London, England
  • London, gb

Neighbourhoods

Neighborhoods are challenging to search for because they are "fuzzy". It is often not clear exactly where the boundary between "Sommers Town, London" and "Camden, London" is.

Ukraine

See also page about Administrative divisions of Ukraine.

Addresses

Postal codes are 5-digit numbers, however they are optional, so rarely used. Usual address patterns are "Street, address, city, region", but reversed order is also used ("Region, city, street, address"). Address is always typed as "Street address" or "Street, address".

Examples:

  • Печерський узвіз 19, Київ
  • Печерський узвіз 19, Київ, Україна
  • вул. О. Дашкевича, 39, Черкаси, Черкаська область, Україна
  • просп. Энтузиастов, 11, Днепрорудное, Запорожская область, Украина, 71630
  • Pecherskiy Uzviz 19, Kyiv, Ukraine

Abbreviations for types of street are often used. Typical abbreviations are "вул." (street), "просп." or "пр-т" (avenue), "пров." (lane), "бульв." or "бул." or "б-р" (boulevard).

States (oblasts)

Oblasts are used for pointing towns, cities which are not capitals of oblasts.

Examples:

  • Житомирська обл.
  • Київська область
  • м. Бердичів, Київська область
  • Zhytomyr oblast
  • Berdychiv, Zhytomyr oblast

"обл." is abbreviation for "область" (oblast).

Neighbourhoods (rajons)

Rajons are used for pointing villages, towns, cities which are not capitals of oblasts.


USA

Addresses

  • 1370 Willow Road, CA
  • 1370 Willow Road, California
  • 1370 Willow Road, Menlo Park
  • 1370 Willow Road, Menlo Park, CA
  • 1370 Willow Road, Menlo Park, CA 94025
  • 1370 Willow Road, Menlo Park, CA 94025-1516
  • 1370 Willow Road, Menlo Park, California
  • 1370 Willow Road, Menlo Park, US
  • 1370 Willow Road, Menlo Park, USA
  • 1370 Willow Road, Menlo Park, united states
  • 1370 Willow Road, Menlo Park, California, US
  • 1370 Willow Road, Menlo Park, California, USA
  • 1370 Willow Road, 94025
  • 1370 Willow Road, 2nd Floor, Menlo Park, CA 94025
  • 1370 Willow Road, 2nd Floor, Menlo Park, CA 94025-1516
  • Willow Road & Hamilton Ave, Menlo Park, CA 94025
  • Willow Road & Hamilton Ave, Menlo Park, CA 94025-1516
  • Willow Road and Hamilton Ave, Menlo Park, CA 94025
  • Willow Road and Hamilton Ave, Menlo Park, CA 94025-1516

Addresses for apartments have: street-number street, # or #appartmentnumber, city, state, zip code

  • 350 Sharon Park Dr, #Z97, Menlo Park, CA 94025

Grid system

The street address component can take two common forms for grid-system street names, as commonly found Utah (eg. Pleasant Grove). The following examples both refer to a point along the way tagged with name="South 600 East" or name="S 600 E" — it is up to the geocoder to determine that "S" and "E" may be equivalent to "South" and "East" in both OpenStreetMap data and in user searches. Many ways imported from TIGER have an abbreviated name, eg. "N 100 W", whereas ideally (and many user-tagged ways) use the full form, eg. "North 100 West".

  • 80 S 600 E, Pleasant Grove, UT
  • 80 South 600 East, Pleasant Grove, UT

Similarly there are named streets with a directional prefix, for example:

  • West Center Street, Pleasant Grove, UT
  • W Center St, Pleasant Grove, UT

Finally, it is possible that a user may refer simply to "Center Street", which is represented in OpenStreetMap by the combination of "West Center Street" and "East Center Street" (eg. Provo). The ideal location would be a node shared by both ways, indicating the logical center of Center Street. This search is less likely than those above as it refers only to a street, not a specific street address.

  • Center Street, Provo, UT

States

  • STATE, ALIAS
  • Arizona, AZ
  • California, CA
  • Nevada, NV
  • Utah, UT

Cities

  • Menlo Park
  • Menlo Park, CA
  • Menlo Park, California
  • Menlo Park, US
  • Menlo Park, united states
  • Menlo Park, California, US