Building Construction Investment by Type of Structure

Four rows from the extract, exactly as they land in a delivery:

REF_DATE: 1959-01   GEO: Canada            Housing_estimates: Housing starts
Type_of_unit: Total units           UOM: units     COORDINATE: 1.1.1
VECTOR: v15817836                   VALUE: 14776

REF_DATE: 1959-01   GEO: Canada            Housing_estimates: Housing starts
Type_of_unit: Single-detached units VALUE: 7087

REF_DATE: 2026-04   GEO: British Columbia  Housing_estimates: Housing starts
Type_of_unit: Row units             VALUE: 768    VECTOR: v15817824

REF_DATE: 2026-04   GEO: British Columbia  Housing_estimates: Housing starts
Type_of_unit: Apartment and other units  VALUE: 8095

One row per geography per measure per dwelling type per quarter. The pair above brackets the whole series: Canada's first recorded quarter - January 1959, 14,776 total starts of which 7,087 were single-detached - sits opposite British Columbia's spring 2026 split, where apartment-and-other starts outnumber row units better than ten to one. Identical columns throughout, so a 1959 recession call and a 2026 provincial forecast load from one schema.

Get a sample of this dataset and we route rows scoped to your province list, dwelling types or date range the same day.

The raw extract carries fifteen columns; the ten below do the analytical work, each definition verified against the delivered file itself. Everything else - scalar factors, symbol markers, decimal conventions - folds out on request with your sample.

Geography - Canada as a national aggregate plus all ten provinces, Newfoundland and Labrador through British Columbia. Territory members exist in the geography dimension for northern analysis, though the historical depth concentrates in the provinces. The full industry context lives on our construction engineering data hub.

Temporal - quarterly reference periods from Q1 1959 through April 2026: sixty-seven years, roughly 270 quarters for every series combination. Few construction series anywhere run that long at provincial granularity, which is why this cube keeps resurfacing in cycle research.

Granularity - province crossed with three housing measures and five dwelling types, restricted to urban centres of 10,000 or more population. Rural construction sits outside the survey frame, worth remembering when a provincial total trails a permit ledger.

API, files, or your warehouse. Daily, weekly, or hourly.

Pick the channel your team already works in: REST endpoints for pulling a single VECTOR-id series or a filtered slice, flat files sized for batch loads (the full extract runs 39,490 rows and expands to 5.8 MB), or a direct pipe into Snowflake, BigQuery or Redshift. Cadence is yours to set - and to change when your models change.

Every delivery ships with the full field dictionary, sample rows for validation, and a schema that holds steady between refreshes.

• Construction-cycle modeling - sixty-seven years of quarterly starts, under-construction and completions give models enough cycles to separate secular trend from rate-driven noise; the workflow is detailed on our data scientists use cases page.
• Provincial demand forecasting - building-products, equipment and trades capacity planners read the province-by-dwelling-type matrix as forward demand: units under construction today are finishings and fittings orders next quarter. More applications on the market researchers use cases page.
• Backtesting homebuilder exposure - quants align provincial construction momentum against listed builder and materials returns across several full cycles, as covered on the investors and quants use cases page.
• Citable trend reporting - journalists and academics get an official national statistical series whose every figure arrives with estimation flags attached, so charts can be sourced and caveated in one step.
• Expansion timing - competitive-intel teams correlate competitor capacity announcements with provincial construction momentum; the pairing is sketched on the competitive intel product teams use cases page.

Data scientists and ML engineers get the highest-leverage read: a sixty-seven-year quarterly panel with verified definitions and stable series IDs, long enough to train across multiple rate regimes. Market researchers and consultants get province-by-dwelling-type demand matrices for sizing studies. Journalists, academics and students get an official series they can cite with quality flags doing the caveating. Investors and quant researchers get a macro overlay for construction-exposed positions. Developers building data products get a schema that has survived decades of revisions intact. Sales and growth teams get province-level momentum signals for territory planning. Competitive intelligence teams get the construction backdrop behind rival capacity moves. Builders wiring the feed into products will find the patterns on the developers builders use cases page.

Within the catalog's construction shelf, this is the deep-history, high-granularity volume play. Value of Construction Put in Place (VIP) answers the money question for the United States - $2,166,539 million SAAR of total construction work in June 2026 - while this cube answers the volume question for Canada; the head-to-head continues in Value of Construction Put in Place vs Building Construction Investment by Type of Structure.

New Residential Construction (Building Permits, Housing Starts & Completions) is the American monthly analogue with permits added, and FRED Economic Data (Housing Starts & Construction Series) carries US starts back to January 1959. Both score 9 on the catalog rubric against this table's 8; neither matches province-level splits reaching back to Q1 1959.

The naming mismatch bears repeating for buyers: if the job is dollar-value construction investment, VIP is the right shelf. If the job is structural volume - who is building what kind of housing, where, since Eisenhower - this cube has no peer at that depth. Volume and dollar readings disagree at turning points, which is why holding both is standard practice. The full shortlist sits on our best construction engineering datasets page.

Four things worth knowing upfront.

First, the label stretches. This record is titled building construction investment by type of structure, but the cube measures dwelling units - starts, under construction, completions - not investment dollars. If you need the money side, say so in the sample request and we will scope the matching dollar series alongside.

Second, geography stops at provinces plus territories, and the survey frame covers urban centres of 10,000 or more population. Rural and small-market construction is out of frame, so totals will trail permit ledgers that count everything a municipality issued.

Third, revisions happen and are flagged rather than hidden: a correction restated September 2023 and Q3 2023 values for affected series, and every cell carries its estimation status. Pin your extract vintage if you re-run comparisons across releases.

Fourth, the cadence is quarterly, so fast-moving questions - this month's momentum - want a monthly complement. Where this cube wins is depth and provincial structure, not speed.

Get a sample of this dataset and we will route rows scoped to your provinces, dwelling types and date range, with the field dictionary and quality-flag legend attached.