Geocoding

Expand all | Collapse all

Avoiding garbage: clean and prep your data

  • 1.  Avoiding garbage: clean and prep your data

    Pitney Bowes
    Posted 10 days ago
    Edited by Eileen Potter 10 days ago

    Avoiding garbage: clean and prep your data

    We all know the old saying "garbage in, garbage out" but what does that mean for your geocoding data? Using complete input addresses that are based on the country's postal authority standards may lead to better matches and/or better locations in the geocoded output.

    What tips or tricks have you incorporated when you prep your data? Feel free to click "Reply" and add your comments.

    Common guidelines include:

    • Input address lists should be as complete as possible based on postal authority standards for the country. Review for misspellings, leading or trailing spaces, and postcodes consistent with the named City.
      • Tip: The Universal Postal Union website is a good source for finding addressing standards by country.
        Make a selection in their "Addressing area" and click Display. The corresponding PDF provides a link to the country's postal website, describes the expected format, and includes some addressing examples.
        Postal standards by country
    • When doing multi-country geocoding, be sure to use the three-letter ISO country code, defined by the International Organization for Standardization (ISO).
    • Sort your input table by postcode before geocoding to help with performance.
    • For single-line input, punctuation is generally ignored. However, using separators (commas, semicolons, etc.) between different address elements may improve results and performance.

    Examples

    The examples below demonstrate forward geocoding from an address, reverse geocoding from x,y coordinates, and multi-line input. They contain both an optional header line and optional input fields.
    • Forward geocoding from an address

      mainAddress;areaName3;areaName1;postalCode;country
      26 EDISON DR;NISKAYUNA;NY;12309;USA
      92 MCCRACKEN RD;MILLBURY;MA;01527;USA
    • Reverse geocoding from x,y coordinates

      X|Y|Country
      -113.532|53.5421|CAN
      -115.423|32.6561|MEX
    • Multi-line input record

      Id|inAddressLine1|inCity|inStateProvince|inPostcode|inCountryName|inCountry_ISO3
      1|SALURNER STR 15|INNSBRUCK||6020|AUSTRIA|AUT
      2|10751 WESTMINSTER DR|SOUTHWOLD|ON|N0L2G0|CANADA|CAN




    ------------------------------
    Eileen Potter
    Knowledge Community Shared Account
    Shelton CT
    ------------------------------


  • 2.  RE: Avoiding garbage: clean and prep your data

    This message was posted by a user wishing to remain anonymous
    Posted 10 days ago
    This post was removed


  • 3.  RE: Avoiding garbage: clean and prep your data

    Posted 7 days ago
    Hi Eileen,

    Your post states: Sort your input table by postcode before geocoding to help with performance.

    Could you explain how/why this works?

    Thanks, Tim


    ------------------------------
    Timothy Mashford
    Melbourne, Australia
    ------------------------------



  • 4.  RE: Avoiding garbage: clean and prep your data

    Pitney Bowes
    Posted 7 days ago

    Hi Tim,

    Sorting by postcode before geocoding can help performance when geocoding a large batch of address data because it can improve reference data search time.

    As part of the geocoding process the geocoder will match the input address to the geocoder reference data by searching reference data files by country, town and\or postcode.

    If the input address is sorted by country and postcode too, the searching of referencing data will be more efficient.

    Did I help answer your question?

    Regards

    Jeff



    ------------------------------
    Jeff Rueger
    PITNEY BOWES SOFTWARE, INC
    Troy NY
    ------------------------------



  • 5.  RE: Avoiding garbage: clean and prep your data

    Posted 4 days ago
    Thanks Jeff, so it's not specific to postcode, but rather any boundary e.g. country, locality/suburb, postcode....

    Good to know, I'll give it a try.

    Regards, Tim

    ------------------------------
    Timothy Mashford
    Melbourne, Australia
    ------------------------------