Method and data

How this was measured

Summary

  • 311 data: every non-duplicate request made Oct 1, 2024 to Sept 30, 2026 (City of Calgary open data). Reporting rates use 30 common service types. "Infrastructure" means 21 request types about roads, water and sewer, and parks and trees.
  • Residents: 2021 Census population by community. Communities with under 500 residents, or created after 2021, are left out, as are Carrington and Livingston (6,735 residents), which fall outside the 2021 Equity Index areas.
  • Community characteristics: the City's Calgary Equity Index (2014–2021 data, measured for areas that group several communities), 2021 Census households and employment by community, 2026 property assessments, building permits and the City's historical community populations.
  • Like-for-like estimate: population-weighted regression on the log of reports per resident, controlling for income, renters, community age and distance from downtown. We ran the main model plus 13 alternatives, and a 1,000-draw bootstrap that resamples whole areas.
  • Similar communities: "expected" reporting comes from a model fitted to all 195 communities using median income, share of renters, distance from downtown and community age. Each community profile also lists its five most similar real communities on those four factors, with a similarity score from 0 to 100 (100 = identical; each 0.1 standard deviation of average difference takes off 5 points).
  • Limits: this shows a pattern, not its cause. Fewer reports could partly reflect fewer problems. The City also finds some problems through its own inspections (for example potholes and sidewalks); the study measures who reports, and City inspection data could show how many problems are caught without a report. Language figures are for Census areas, not individual residents, so the data can't show that the residents who don't speak English are the ones not reporting. Most records locate requests only to the community's centre point. Population is from the 2021 Census, the latest count by community. Eight communities have grown a lot since (an estimated 1.5 times their 2021 population or more, from homes on the property assessment roll and rental-building permits), so their per-resident rates are overstated and their reporting gap is not shown; profiles of every community that grew 15% or more also show growth-adjusted figures. Growth was lowest in the communities with the most non-English speakers, so it doesn't explain the gap.

Downloads

Re-run the study yourself

Quickest check (about 2 minutes): download the study in one Python script, install Python 3 from python.org, and run python who_calgarys_311_hears.py (on Windows, py who_calgarys_311_hears.py, or double-click it). It needs nothing else: it installs pandas and numpy if they are missing, downloads 8 data sets from Open Data, and prints every main result in 12 steps next to the published figures: the communities, the raw gap, communities of similar age, income and language side by side, the fair-comparison model, 1,000 bootstrap re-runs, every request type, response speed, app use and Ward 5. To see the code one step at a time, use the notebook in the full study package (folder 00_Step_by_step): open it in Jupyter or Google Colab (File, Upload notebook) and run each block in turn.

The full study (about 15 minutes), with all 14 model versions and the bootstrap range: anyone can rebuild it from City of Calgary Open Data. You need Python 3.10 or newer and an internet connection. Unzip the full study package, open a terminal in its folder and run (on Windows, type py instead of python):

  1. python -m pip install pandas numpy (once per computer)
  2. python download_data.py downloads all 16 source extracts from Open Data (about a minute)
  3. python build_inputs.py prepares them for the analysis
  4. cd 02_Analysis, then python reanalyze.py runs the analysis (2 to 3 minutes), then cd ..
  5. python check_results.py compares your results with the published ones

Skip step 2 to reproduce the published figures exactly from the extracts in the package (downloaded October 20–21, 2026). A fresh download gives the same findings, with small differences in counts because Open Data is live: requests are closed, re-coded or merged as duplicates after the fact. Checked on October 31, 2026, every finding was identical, with 809,532 requests instead of 809,528. Full instructions are in the package's README.txt.

About this study

This is an independent analysis of public City of Calgary data. It isn't an audit of any person, department or contractor, and it doesn't allege wrongdoing. The aim is to help the City, Council, community associations and residents make 311 work equally well for everyone.

Every figure was independently recomputed from the raw data, and the errors found were corrected. If you spot a problem, please report a correction.

Sources: 311 Service Requests · Calgary Equity Index, 2021 Census by community, Current Year Property Assessments, Building Permits, Historical Community Populations (City of Calgary open data) · 311 Calgary app · Related research: Kontokosta et al. (2017), New York City. Not affiliated with The City of Calgary.