Methodology

The Planada Score

v4.1 · last updated May 11, 2026

What it is

A 0–100 quality-of-life composite for every Census Subdivision in Canada — 5,161 places. Computed from ~40 underlying indicators across 9 dimensions. The score is one lens, not a definitive ranking.

The 9 dimensions

Seven dimensions synthesize three established quality-of-life frameworks: Statistics Canada's Quality of Life Framework, the Canadian Index of Wellbeing (University of Waterloo), and AARP's Livability Index methodology. We added two more — Livability and Opportunity — to capture daily-life amenity access, environmental risk, and economic dynamism that the traditional 7 don't reach.

  • Prosperity (14.3%) — income, employment rate, industry diversity, income inequality
  • Housing (14.3%) — renter shelter burden, owner shelter burden, rent-to-income, dwelling-to-income, core housing need
  • Equity (11.8%) — low-income share, Gini coefficient, settled-immigrant share
  • Safety (10.9%) — Crime Severity Index, Violent CSI, Non-Violent CSI
  • Education & opportunity (10.9%) — bachelor's+ share, post-secondary share, employment access
  • Mobility (10.9%) — short-commute share, sustainable-mode share (transit, walk, bike), long-commute share inverted
  • Community vitality (10.9%) — owner share, family-household share, multigenerational share
  • Livability (8.0%) — proximity to health, parks, grocery, libraries (StatsCan PMD); cultural density per 10,000 residents (theatres, cinemas, museums, arts centres from OpenStreetMap); wildfire risk (NRCan National Fire Database, 2014–2023, inverted)
  • Opportunity (8.0%) — knowledge economy share (NAICS-derived), federal voter turnout, recent population change, business demography (active businesses per capita, opening rate, net creation rate from StatsCan 33-10-0270), interprovincial migration + non-permanent resident inflow (StatsCan 17-10-0149, CMA-fanned to constituent CSDs)

The math

Each indicator is processed through a four-step pipeline before it contributes to a dimension score:

  1. Log-transform indicators with skewed distributions (income, density, cultural density per capita) so a few extreme values don't dominate.
  2. Percentile-rank against the national CSD pool (3,995 regular CSDs). Indian Reserves, Northern Regions, and Settlements are excluded from the percentile pool to avoid distortion from systematic data suppression.
  3. Outlier capping at the 1st and 99th percentile so single extreme values cannot skew the percentile distribution.
  4. Sign inversion for lower-is-better indicators (crime, unemployment, shelter burden, wildfire risk) so higher value always means better score.

Within each dimension, indicators are combined with inverse-variance weighting — indicators with higher national variance get less weight. This keeps noisy signals from dominating the dimension score.

The composite is the framework-derived weighted average of the 9 dimensions, mapped to 0–100. Buckets are calibrated to the actual score distribution across all 5,161 places (median 50.8, range 11.8–84.3): STELLAR ≥ 65 (top ≈ 12%) · STRONG ≥ 50 (above median) · FAIR ≥ 35 · WEAK < 35 (bottom ≈ 10%). Each drawer also shows a national percentile — "above X% of Canadian places" — so the absolute number reads in context.

Bracketed mode

In Compare and elsewhere, you can switch to bracketed scoring. Same formula, but indicators are percentile-ranked within a population class — Rural / Small town / Mid-size / Large city / Metro. Lets a mid-sized Quebec town be compared to peers instead of competing with Toronto on density. Both modes are honest. The national score answers "where does this place sit nationally?"; the bracketed score answers "how is this place doing relative to similar-sized communities?".

Rankings (Top 10 / Bottom 10)

The Top 10 / Bottom 10 lists in the side panel filter out two kinds of places to keep them honest:

  • Places with INSUFFICIENT DATA — no honest score, so they shouldn't appear in any ranking.
  • Places with population < 5,000 — micro-CSDs (Pangman 60, Hill Spring 190, dozens of similar rural shells) produce statistical-artifact rankings from thin data. The pop floor keeps lists about real cities.

The map itself shows every CSD with sufficient data — the filter applies only to ranked lists.

Data sources

  • Statistics Canada — 2021 Census of Population (≈ 70% of indicators)
  • Statistics Canada — Labour Force Survey (monthly, CMA-level for major metros)
  • Statistics Canada — Uniform Crime Reporting Program (annual, CMA + province level)
  • Statistics Canada — Proximity Measures Database (Livability — 4 proximity indicators)
  • Statistics Canada — Table 33-10-0270 (business openings/closings, monthly seasonally adjusted)
  • Statistics Canada — Table 17-10-0149 (components of population change, 2021 boundaries)
  • OpenStreetMap (Overpass API) (Livability — 4 cultural density indicators, joined to StatsCan 2021 cartographic CSD boundaries)
  • NRCan Canadian National Fire Database (Livability — 2 wildfire risk indicators, 2014–2023 window)
  • Elections Canada (federal turnout by riding, mapped to CSD)
  • Statistics Canada — Geographic Attribute File (boundaries, CMA assignment, population brackets)

Each indicator carries a freshness label in the UI. Green pills mark monthly LFS / business demography. Neutral chips mark 2021 Census values that haven't refreshed since the census cycle.

Coverage gates

A region's score is suppressed and shown as INSUFFICIENT DATA when fewer than 5 of the 9 dimensions have at least 50% of their indicators populated. This affects ≈ 1,400 of 5,161 places — mostly small CSDs where StatsCan suppresses for confidentiality, Indian Reserves with systematic data suppression, Northern Regions with thinner survey coverage, and unincorporated Settlements.

The two new dimensions (Livability, Opportunity) have systematically thinner coverage than the original seven — PMD doesn't score every dissemination block; OSM cultural density is suppressed for places under 1,000 residents; CMA- level indicators (business demography, migration) only fan out to constituent CSDs of metropolitan areas. The completeness gate is calibrated to admit places with strong traditional census coverage even when the new dimensions don't reach them.

Indian Reserves and Indigenous communities

Indian Reserves are flagged separately on the map and in the drawer. Statistics Canada systematically suppresses census variables on reserves for sample-size and confidentiality reasons, and First Nations measure community wellbeing through their own frameworks — including the FNIGC's Regional Health Survey and the OCAP principles for data sovereignty. The Planada Score's methodology is designed for the broader Canadian census ecosystem and is not a fair lens for Indigenous communities. Where reserve data is shown, it is partial by design.

What we don't measure

Nine dimensions still leave gaps. Notably absent:

  • Air quality (ECCC NAPS PM2.5 / NO₂ — pending data ingestion)
  • Flood and heat risk per address (NRCan flood maps + ClimateData.ca grid — pending GIS joins)
  • School quality (we proxy with attainment, which is weak)
  • Healthcare wait times (uneven provincial reporting)
  • Income mobility (StatsCan IMDB — separate access pattern)
  • Subjective wellbeing (no Canada-wide standardized survey)

These gaps mean the score under-rewards big cities with strong cultural, environmental, and amenity life that the current indicator set doesn't capture, and over-rewards small cultural touristy towns with high per-capita amenity counts. We add data as we ingest it cleanly — dimensions expand rather than shift, so historic scores stay roughly comparable.

Limitations

  • Most Baseline indicators are from the 2021 Census, ≈ 4 years old. Only Labour Force Survey rates, business demography, wildfire counts, and OSM amenities update on a faster cadence.
  • Equal weighting of indicators within Livability and Opportunity (after percentile transformation) is an editorial choice. Reasonable people may prefer different mixes.
  • Dimension weights average citations from three reference frameworks for the original 7; the two new dimensions sit at 8% each as a cautious entry given lighter evidence base. This is editorial.
  • Per-capita normalization of cultural amenities favors mid-size cultural cities (Charlottetown, Stratford) over dense metros (Toronto). The wildfire indicator counter- balances this — small BC and northern AB places take a hit that big eastern metros don't face.
  • A single composite cannot capture the lived experience of 5,161 different communities. The dimension breakdown exists so users can see where a place is strong or weak rather than relying on the headline number.
  • We have not externally validated the score against resident-reported life satisfaction.

Corrections

If you find an error or want to challenge a score, write to [email protected]. We aim to respond to evidence-based corrections within 48 hours. Public corrections are logged.

Version history

  • v4.1 (May 11, 2026) — Bucket thresholds recalibrated to the actual score distribution (STELLAR ≥ 65 from 75, STRONG ≥ 50 from 55) so metros like Calgary read as STRONG rather than FAIR. Added national percentile readout — "above X% of Canadian places" — to the drawer for context.
  • v4.0 (May 9, 2026) — Score expanded from 7 to 9 dimensions. Livability (proximity + cultural density + wildfire risk) and Opportunity (knowledge economy + business demography + migration + civic engagement) added at 8% each, original 7 rescaled to sum to 84%. Top 10 / Bottom 10 lists now exclude INSUFFICIENT_DATA and require pop ≥ 5,000. The earlier v3 three-ring system was retired in favor of one composite — same data, single-number outcome.
  • v3.x (May 2026) — Three-ring experiment (Baseline + Livability + Opportunity as separate composites). Retired after community feedback that one number per place is the right primary view.
  • v2.0 (Apr 2026) — 7-dimensional composite with inverse-variance weighting, outlier capping, data-completeness gates, separate Indigenous community framing.
  • v1.0 (Apr 2026) — deprecated. 5-component composite (income, education, growth, ownership, density). Over-rewarded wealthy suburbs and didn't weight housing affordability.