How OpenStats reads Statistics Canada building permits

Construction Pulse and the OpenStats permit rankings are not a dump of every row in the permits atlas. They use Statistics Canada Table 34-10-0292-01 rows tagged as direct live readings. Anything marked seed, regional proxy, or draft stays on city pages for coverage but stays out of national totals and acceleration tables. Understanding that gate is the key to reading May 2026 headlines like the $9.5B national total or Peterborough’s +1,653% acceleration line.

Why the gate exists

A missing CMA table should not invent a false national spike. Mixing proxy estimates into a Canada total would also make month-over-month changes impossible to interpret. The live gate keeps Pulse comparable across months even when atlas coverage is wider than the StatCan direct set. Atlas pages may show a seed row so a reader in an uncovered town still has a labeled page; Pulse and the permit value ranking simply ignore that row when summing national or calendar-year leaderboards.

Direct rows come from the StatCan CMA monthly building permits release. OpenStats ingests value, count, and residential/non-residential splits where the table publishes them. When StatCan revises a prior month, OpenStats replays the revision on the next pipeline run and notes the change on the city page. Seed rows, by contrast, are static placeholders until a live feed replaces them.

What readers get on each surface

• Construction Pulse: national totals, residential/non-residential shares, top value cities, top accelerators (prior-year base ≥ $5M).
• Permit value ranking: calendar-year sum of direct readings. Toronto led at about $11.7B YTD in the current edition, Vancouver at about $9.5B, Montreal at about $5.3B.
• Permits atlas: per-city history, including labeled non-direct rows where we still publish a page.

May 2026 illustrates the split. National direct-tier value was about $9.5B. Toronto filed about $1.99B in that month alone. Calgary ($579M) and Edmonton ($533M) counted because both carry live CMA rows. A city with only a seed row might still have an atlas page for reader convenience, but its dollars would be absent from the Pulse national sum until the live feed lands.

Acceleration rules

Pulse acceleration requires at least $5M in the same month one year earlier. Peterborough’s +1,653% line paired $269M in May 2026 with about $15M in May 2025. Ottawa’s +121% paired $399M with about $180M. Without the floor, thin-base percentages would dominate the news table. The rule is published on the Pulse report and repeated here so desk stories stay aligned with the PDF-style report page.

Residential share columns on Pulse city tables show where housing dominates. Calgary’s May split was about 78% residential ($450M of $579M). Edmonton’s was about 69% residential ($369M of $533M). Those splits come from the same direct StatCan rows as the headline values.

Caveats and linked reading

Full caveats live on methodology, named tables on data sources, and correction policy on editorial standards. For a worked example with May 2026 dollars, open the national Pulse story. For Alberta monthly splits, open the Calgary-Edmonton May story. For YTD leaders, open the 2026 permit-value leaders story. Tier labels on every atlas slug remain the ground truth for whether a number enters Pulse math. Ottawa’s May 2026 permit value of $399M (+121% YoY) and Peterborough’s $269M (+1,653% YoY) both use the same direct-tier gate as Toronto’s $1.99B row. Vancouver about $1.18B and Montreal about $1.08B in May 2026 likewise enter Pulse only when their CMA rows are tagged direct in Table 34-10-0292-01. Edmonton about $533M and Calgary about $579M in May 2026 entered the same national $9.5B sum under identical tier rules as the three largest metros. National May residential share columns on Pulse help readers see whether the $9.5B month was housing-heavy or carried by institutional filings in individual CMAs. Calendar-year permit value ranking sums use the same direct-tier filter as Pulse, which keeps Toronto about $11.7B YTD and Vancouver about $9.5B YTD on the same footing as monthly Pulse rows.