Canada Data Release Notes
Updated over a week ago


This page contains release notes for each of Lightcast’s bi-annual Canada dataruns from the 2018.3 datarun forward. The release notes contain information on major methodology changes included with each datarun.

  • The Upcoming Changes section outlines major changes users can expect to see in the next datarun.

  • The Current Release Notes section lists the changes and updates introduced with the latest datarun available in Analyst/Developer.

  • The Dataset Chart shows a list of the major sources that go into Lightcast data and what "vintage" (year, month, etc.) of each source was used for each of the past four dataruns.

  • The Older Release Notes section contains release notes from prior dataruns.

To read more on Lightcast data updates, see this article.

Expected release dates for the next two dataruns

  • 2024.3: October 2024

  • 2024.1: May 2025

Latest Release Notes

This section lists the changes and updates introduced in the latest datarun.

See current Dataset Chart here


beta release: May 16, 2024

Final release : May 28, 2024

Note: On June 3, 2024, Lightcast updated Canada datarun 2024.1 to address a bug in the data. The bug had led to incorrect 2023 December business location counts. Correcting that error also influences our employment estimates, primarily in government and agriculture, across the country.

Data Updates:

No major classification updates or methodology changes.

Older Release Notes


beta release: October 3, 2023

Final release : October 10, 2023

Data Updates:
As part of the 2023.3 Datarun the following have been updated:

  • Canadian Business Counts

  • Survey of Employment, Payrolls, and Hours

  • Symmetric Input-Output Tables

Revised Mappings Between Census Subdivision(CSD) Versions

We revised our mappings between Census subdivision (CSD) versions. There are several cases in which tiny sections of CSDs were broken off to form new CSDs or to become part of other CSDs, and our old mappings weren't dealing with these appropriately. This caused problems especially in rural parts of Canada, in which very small transfers of land between CSDs (sometimes a transfer of as little as a few hundred square yards) caused large shifts in employment between the two CSDs. With the new mappings, this problem no longer exists. The instances of this we found were corrected, and we continue to try to track down any additional instances of this problem.


beta release: June 15, 2023

Final release est.: June 21, 2023

Update to NOC 2021

Lightcast's 2023.1 Canadian datarun incorporates the latest version of occupation codes (NOC 2021). NOC 2021 represents a substantive break from previous NOC versions in that the most granular categories are represented by 5-digit codes rather than by 4-digit codes as was the case with previous releases.

Census 2021 Update

The 2023.1 release incorporates data from the 2021 Census, which has a substantial effect on our employment estimates. Occupation data in the census was released in the NOC 2021 classification system, and the geographic system in this release has been updated to Standard Geographic Classification (SGC) 2021 to correspond with the census release as well.


None. There were no classification updates or major methodology changes.


Update to CIP 2016

We will now be using CIP 2016 to categorize PSIS educational attainment data. More specifically, we will be using the variant of CIP 2016 created for the 2018 release of PSIS that includes cannabis-related programs (see )[


Revised staffing processWe are now using year-specific seeds (where possible) for the staffing unsuppression process. These new seeds come from a linear interpolation of census staffing data from 2006 (the first Census staffing year) to to the latest Census staffing year. Unsuppression of data in earlier years now uses seeds more heavily on the 2006 Census, and later years use seeds based more heavily on the latest Census. This replaces the previous method in which all seeds came from a simple averaging of 2006 and latest-year data. Note that we do not include 2011 census data in the linear interpolation; data from 2011 is unreliable because the 2011 National Household Survey had a very poor response rate (69% vs 94% of 2006 Census).

Fewer tiny industry employment valuesWe removed some abnormally small employment values (less than one job) from Lightcast industry data and distributed them to other industries with more reasonable employment values, effectively clustering the data more tightly around known values. The effect on industry, staffing, and occupation data will make the resulting data more reasonable and queries faster, and will likely be unnoticeable because of routine data source updates.

SEPH monthly averagingPreviously, our methodology only allowed us to use monthly SEPH data once all twelve months had been released. This led to accurate data, but we lagged behind the latest SEPH releases for most of the year. We’ve changed this methodology to all us to use the average of the most recent twelve months of monthly SEPH data as the estimate current year industry employment in cases where a full calendar year of monthly SEPH data isn’t available yet. To help keep the data consistent, we will only do this once we have more than six months of data for that year. So, for example, once we have SEPH monthly data through July of 2022, the average of August 2021 through July of 2022 will be released as 2022 data.[/su_spoiler][su_spoiler title="2021.1" style="fancy"]


Revised PSIS education data

With the 2018/19 release of the Postsecondary Student Information System (PSIS) data, PSIS stopped providing all-institution totals (summing data across all institutions). This change affects all years of historical data and has resulted in slightly different results while de-rounding the data (PSIS rounds to the nearest 3). Additionally, PSIS began including data for Cannabis-related instructional programs (CIP code 71).

Revised LFS Industry and Occupation Employment DataThe most recent release of Labour Force Survey industry and occupation employment tables included re-based employment values from 2006 forward. This re-base was done to ensure that the data reflect the most recently available population estimates (based on the 2016 Census), 2016 geographic boundaries, and more recent industry classifications. As a result, employment and self-employment values for 2006-2019 will differ slightly from previous estimates.As mentioned above, the Labour Force Survey industry classification has been updated from using NAICS 2012 to NAICS Canada 2017 version 3.0. As a result, Lightcast’s industry self-employment data now includes the following two Cannabis-related industries:

  • 3123 - Cannabis product manufacturing

  • 4134 - Cannabis merchant wholesalers

COPS Occupation Projection Adjustment BugWe found and fixed a small bug in our COPS projection adjustment where occupations were being adjusted to aggregate 3-digit NOC COPS projections instead of to 4-digit NOC COPS projections.

Revised Population Demographics DataThe most recent release of some of Statcan’s demographics tables included data for all years from 2001 onward. Before this, those files only contained years from 2006 onward, so we used archived data to get 2001-2005 demographics. We’ve now switched to only using the actively released tables, but data for 2001-2005 have been revised, so they differ very slightly from the data that existed in the archived dataset we were using. For more information about the tables in question, see and


Improved PSIS DetailPSIS data granularity has been significantly improved. CIP detail has been expanded from the 4 to 6-digit CIP level. In addition, PSIS data now also includes a Credential dimension, which is the level of education previously achieved by the student. For more information about this dimension and a description of the codes in this dimension, see: Note, the Credential dimension should not be confused with the Award Level dimension, which is the context in which a program is offered (i.e. the level of the studies and whether the student is seeking the degree as preparation for a career or for further studies). For more information about this dimension and a description of the codes in this dimension, see:


COPS Projection Adjustment Bug

We found and fixed a small bug in our COPS projection adjustment that caused our adjusted projections to be up to 6% higher than COPS projections in the final projection year. Lightcast now matches projection trends from COPS.

Use Beginning-of-Year Minimum Wage

Provinces and Territories sometimes increase their minimum wage multiple times per year. Lightcast has switched from using end-of-year to beginning-of-year minimum wage laws as the wage floor for each year’s occupational earnings.

Correct CIP Code Titles

We fixed the titles of the following CIP codes:CIP 11: changed from “Communications technologies/technicians and support services” to “Computer and information sciences and support services”CIP 26: English title changed from “Sciences biologiques et biomédicales” to “Biological and biomedical sciences”

Industry Earnings Estimation

We updated our earnings estimation methodology for SEPH, generally making earnings per job more reasonable for Canadian industry data. Earnings did not change significantly.

Industry Unemployment Data

We added unemployment data by 2-digit NAICS, from the Labour Force Survey. This data will be available in the 2020.1 and following dataruns. It can be found in the columns options in the Industry Table. For more information on digit level available, see this article.[/su_spoiler][su_spoiler title="2019.3" style="fancy"]


None. There were no classification updates or major methodology changes.


Lightcast Industry Classification (Lightcast NAICS) 2017

Lightcast’s industry classification has been updated to closely match the NAICS Canada 2017 version 3.0 classification, which includes the following new industry codes:

  • 111412 – Cannabis grown under cover

  • 111995 – Cannabis grown in open fields

  • 3123 and 312310 – Cannabis product manufacturing

  • 4134 and 413410 – Cannabis merchant wholesalers

  • 453993 – Cannabis stores

For more information on the changes between NAICS 2017 v2 and NAICS 2017 v3, see StatCan's mapping between them.

While SEPH now covers Cannabis, employment in industries 3123 and 4134 are usually suppressed. SEPH also now suppresses employment in the sibling industries of 3122 - Tobacco manufacturing and 4133 - Cigarette and tobacco product merchant wholesalers. To help unsuppress these values, we estimate employment using Canadian Business Counts (the only other data source we use that covers the cannabis industries) and use those estimates as seed values in our unsuppression.


Standard Geographical Classification (SGC) 2016

Lightcast’s geography classification has been updated to match the geography classification used by the 2016 Census. For more information, see StatCan's writeup.

Lightcast Industry Classification (Lightcast NAICS) 2017

Lightcast’s industry classification has been updated to closely match NAICS Canada 2017 version 2.0 classification. As with Lightcast NAICS 2012, Lightcast NAICS 2017 combines or omits a few of the standard NAICS codes:


NAICS 2017

Emsi NAICS 2017

Farming and Agriculture

111* - 112*


Scenic and Sightseeing Transportation, Other


Rolled into 4871 and 4872

Other Federal Services

9112, 9113, 9114, 9119

Rolled into 9112

Provincial and Territorial Public Administration



Local, Municipal, and Regional Public Administration



International Public Administration


Not Included

For more information on the changes between NAICS 2012 and NAICS 2017 v2, see StatCan's Revision of NAICS Canada for 2017 (Version 1.0) and NAICS Canada 2017 Version 2.0.

National Occupational Classification (NOC) 2016 Version 1.1

Lightcast’s occupation classification has been updated to match NOC 2016 version 1.1. The structure of NOC 2016 Version 1.1 remains unchanged from that of NOC 2011. No major groups, minor groups or unit groups have been added, deleted or combined, though some groups have new names or updated content. For more information, see StatCan's summary of the changes.

LFS Volatility Adjustment

Due to the volatility of historical LFS data, Lightcast is now smoothing historical LFS using Kaufman's Adaptive Moving Average (KAMA). KAMA is based on the Exponential Moving Average (EMA) and is responsive to both trend and volatility. It closely follows employment when volatility is low and smooths out volatility when employment fluctuates dramatically. For more information about this methodology or to request the methodology white paper, please contact your Lightcast representative. [/su_spoiler]

Did this answer your question?