Canadian Housing Observatory Data Repository

This repository serves as the upstream data source for all Canadian Housing Observatory repositories and releases.

Getting Started

git clone git@github.com:mlian031/canhopi.git
cd canhopi
R -q -e 'renv::restore()'

Raw input files are intentionally kept out of version control because some of the source files exceed GitHub's 100 MB per-file limit.

Before running the project, download the relevant data from CREA MLS and Statistics Canada and place them at these exact paths (we will try and get a version that ingests via API up soon!):

data/
  crea/
    MLS_HPI_March_2026/
      Not Seasonally Adjusted (A).xlsx
      Not Seasonally Adjusted (M).xlsx
      Not Seasonally Adjusted (Q).xlsx
      Seasonally Adjusted (M).xlsx
  statscan/
    11100190/
      11100190.csv
      11100190_MetaData.csv
    18100004/
      18100004.csv
      18100004_MetaData.csv
    34100133/
      34100133.csv
      34100133_MetaData.csv

Data Sources

MLS_HPI_March_2026 corresponds to CREA MLS Home Price Index

11100190 corresponds to Table 11-10-0190-01, Market income, government transfers, total income, income tax and after-tax income by economic family type.

18100004 corresponds to Table 18-10-0004-01, Consumer Price Index, monthly, not seasonally adjusted.

34100133 corresponds to Table 34-10-0133-01, Canada Mortgage and Housing Corporation, average rents for areas with a population of 10,000 and over.

Scripts

All analysis code lives under R/.

Run order

Intermediate steps can be run in any order once inputs exist; the final step must be last.

1. R/income.R builds median total income panels (provinces and metropolitan) and saves the intermediate output objects to output/.
2. R/rent.R builds a metropolitan two-bedroom rent panel (2023 real dollars) and saves the intermediate output objects to output/.
3. R/crea.R builds a CREA MLS house price benchmark panels (provinces and metropolitan, 2023 real dollars) and saves the intermediate output objects to output/.
4. R/join_panel.R merges the above into the published provincial and metropolitan panels (and CSV exports).
5. R/bsadf.R computes Backward Supremum ADF (BSADF) paths and generalized SADF (GSADF) test-statistic summaries for each region’s series, using exuber. Requires output/prov_panel.rds and output/metro_panel.rds from step 4.

From the repository root for example:

Rscript R/income.R
Rscript R/rent.R
Rscript R/crea.R
Rscript R/join_panel.R
Rscript R/bsadf.R

Helper scripts

File	Role
R/lookups.R	Shared province_lookup and metro_lookup tables so CREA region names align with Statistics Canada geography labels used in the Statistics Canada tables.
R/cpi_helper.R	build_monthly_cpi_dfltr() for monthly series (CREA) and build_annual_cpi_dfltr() for annual series, using table 18-10-0004-01 (All-items CPI, 2002=100). Real-dollar base year is 2023 in the current code to match with the income data dollars.

Output artifacts

File	Description
output/prov_med_income.rds	Provincial median income (annual).
output/metro_med_income.rds	Metropolitan median income (annual).
output/metro_rent.rds	Metropolitan two-bedroom rent (annual, real 2023$).
output/prov_crea_series.rds	Provincial CREA benchmarks (monthly, real 2023$).
output/metro_crea_series.rds	Metropolitan CREA benchmarks (monthly, real 2023$).
output/prov_panel.rds / output/provincial_panel.csv	Final provincial panel and CSV export, long format.
output/metro_panel.rds / output/metro_panel.csv	Final metropolitan panel and CSV export, long format.
output/bsadf/prov_bubble_paths.rds / output/bsadf/prov_bubble_paths.csv	Provincial BSADF paths, critical values, and exuberance flags (95% level) per date.
output/bsadf/metro_bubble_paths.rds / output/bsadf/metro_bubble_paths.csv	Metropolitan BSADF paths.
output/bsadf/prov_bubble_summary.rds / output/bsadf/prov_bubble_summary.csv	Provincial GSADF statistics and critical values per region.
output/bsadf/metro_bubble_summary.rds / output/bsadf/metro_bubble_summary.csv	Metropolitan GSADF summaries.

Contact

Maintainer: Mike Liang (mlian031@uottawa.ca)
Issues: Report here

Acknowledgements

We thank the Dallas Federal Reserve, Lancaster University Management School (LUMS), International and UK Housing Observatory for inspiring this work.