Question:
Our data team has discovered a rather large discrepancy between the data sets that we received in 2017 and 2018.
In 2017 (25-Aug-2017) we have received the following files:
* pccfNat_AUG15_fccpNat.txt
* pcfrfNatFED2013_AUG15_fcpcefNatCEF2013.txt
In 2018 (25-Jun-2018) we have received the following files:
* pccfNat_fccpNat_062017.txt
* pcfrf_NatFED2013_062017_fcpcefNatCÉF2013.txt
We have discovered that unusually large number of postal codes is not getting a match in our system, that prompted us to compare 2017 and 2018 data files. We have discovered that 16,147 postal codes exist in 2017 data file but do not exist in 2018 data file. We have spot checked 74 postal codes on the Canada Post’s web site and have received positive confirmation for 50 of them, meaning that in 50 cases Canada Post web site would return a list of addresses linked to the given postal code.
To this message is attached an excel file that lists all the postal codes that exist in 2017 file but do not exist in 2018 file.
Answer:
We received the following response from subject matter:
“After doing a little analysis with the provided list of postal code, 10419 postal code from your list are retired. It is important to note that the PCCF based on 2016 geographies does not contain the retired postal code that retired before 2016.
Postal codes are intended for the distribution of mail by Canada Post. The files undergo changes on a regular basis. In some instances, postal codes are retired and re-birthed with a different deliver mode type or the address information for postal codes is changed, which can break our linkage for our geocoding process.
Many postal code did not go through our geocoding process, therefore did not make it to the PCCF file. Every release we are working on getting these postal code back in the product file. The vast majority of these linkages are created in an automated fashion at the dissemination area, dissemination block or block face level geographies. Records that do not link are output for manual geocoding. These records that were previously manually geocoded to the three principle geographies, will now be linked to the census subdivision geography only. They will include only postal codes that do not already appear (already have records) on the main processing table, PCINFO. These records (postal codes) once linked to the CSD level geography, will then be appended to the PCCF and PCFRF product files. They will not be written back to PCINFO, as the processing system (PCUS) will attempt to geocode these postal codes with each subsequent month of processing CPC data.
We are working diligently to continually improve the product with each release quarterly as we are working in improving our production tool. ”
