Methodology

How a community page comes together.

Four steps from public data to editorial summary — and an honest list of what this site is not.

01

Pull

Every month we collect fresh data from the City of Calgary, the Calgary Police Service, Calgary Transit, and the public schools directory. Crime, assessments, schools, parks, services, transit, fire — every feed has a single point of origin we can name.

02

Reconcile

Every record is tagged to an official Calgary community. Edge cases and residual sub-areas are flagged. Rows that disappeared upstream are archived with a reason — never silently dropped — so historical context survives across releases.

03

Visualize

Charts use the same scales across communities so comparisons are honest, not flattering. No red, no green. Direction of change is carried by a glyph and weight, never by colour swap. Every figure shows its source and its date.

04

Narrate

An AI summary describes what the data shows using cautious verbs — appears to, tends to, observed. Every insight is grounded in the same numbers on the page, dated, and labelled. The model speaks like an analyst, not a salesperson.

Where the data comes from

Eight feeds. One destination.

DatasetSourceCadence
Community boundariesCity of CalgaryMonthly
Crime statisticsCalgary Police ServiceQuarterly
Property assessmentsCity of CalgaryAnnual
SchoolsCity of CalgaryMonthly
ParksCity of CalgaryMonthly
Services & amenitiesCity of CalgaryMonthly
Fire stationsCity of CalgaryMonthly
Transit stops & LRTCalgary TransitMonthly
Limitations

What this site doesn't do, and shouldn't.

  • Not real-estate advice. Assessed values are not sale prices. The site won't replace a realtor or a comparative market analysis.
  • Not legal or financial advice. If a number affects a decision, verify it at the source.
  • Not real-time. At worst, one quarter behind for crime, one month behind for everything else.
  • Not exhaustive. Demographics and recent news are on the roadmap but not yet ingested. Side-by-side community comparison is coming soon.
  • Not infallible. AI summaries are grounded in the data shown on the page, but the wording is the model's. Read it as a draft, not a verdict.

Trust is built one footnote at a time.