Gas Bills by Postal Code
Postal code gas consumption screening to identify neighbourhoods with poor building envelopes for municipal retrofit incentive targeting
Key Findings
- ● Postal code gas aggregation carries a usable thermal performance signal for neighbourhood-level screening
- ● Normalization by heated volume from MPAC data is essential — raw slopes reflect building count and size, not envelope quality
- ● Structure type grouping prevents unfair comparison between detached homes and row houses
- ● Top-quartile postal codes represent the strongest candidates for targeted retrofit incentive programs
- ● Ranking is robust to reasonable variation in basement fraction and floor height assumptions
- ● Annual refresh of the screening enables neighbourhood-level M&V of retrofit program impact
Scope
A 3-notebook analytical project that uses aggregated residential natural gas consumption data by postal code, heating degree day regression, and MPAC property tax roll characteristics to screen neighbourhoods for building envelope performance. The target user is a municipal energy planner or community energy organization seeking to direct retrofit incentive dollars to the areas where building envelopes are underperforming most — and where program uptake and savings are most likely. The methodology works with data that Ontario municipalities already have access to: utility-provided gas consumption aggregates and MPAC property assessment records.
Data
Two years of monthly residential natural gas consumption aggregated by postal code (150 postal codes, 24 months, ~3,600 observations), sourced from utility community energy planning data agreements. Regional monthly heating degree days at the 18°C base from Environment and Climate Change Canada. MPAC property tax roll summaries per postal code: average building footprint area, storey count, dominant structure type (detached, semi-detached, row, low-rise apartment), and basement indicator (finished, unfinished, none). The synthetic demonstration dataset is calibrated to reproduce the statistical properties of a mid-size Ontario municipality.
Analytical Approach
For each postal code, an OLS regression of monthly residential gas consumption against monthly HDD produces a thermal slope — the aggregate thermal response of all residential buildings in that area. Quality filters (R² > 0.80, positive slope, reasonable baseload, minimum customer count) remove unreliable regressions. The raw slope is then normalized: first divided by residential property count to get a per-property slope, then divided by estimated heated volume (computed from MPAC footprint, storeys, and basement data) to produce a normalized thermal intensity metric in GJ per HDD per m³. Postal codes are grouped by dominant structure type before ranking, since detached homes have fundamentally different exposed surface area than row houses. The top quartile within each group forms the priority targeting list. Sensitivity analysis confirms ranking stability under normalization parameter variation.
Outcome
The screening identifies the top quartile of postal codes by normalized thermal intensity within each structure type group — the neighbourhoods where building envelopes are losing the most heat per unit of heated space relative to comparable building stock. Normalization materially reshuffles the raw slope ranking, confirming that building stock quantity and size effects must be removed before comparison. The ranking is robust to reasonable variation in normalization assumptions (basement heating fraction, floor height). The targeting table — postal code, structure type, thermal intensity, and priority flag — feeds directly into municipal incentive program design for geographically targeted outreach, audit campaigns, and incentive allocation. An annual refresh cycle enables macro-level M&V of neighbourhood-scale retrofit program impact.
Every municipality running a building retrofit incentive program faces the same question: where should the money go? First-come-first-served allocation suffers from self-selection bias — early adopters in newer neighbourhoods apply first, while areas with the worst envelopes and most to gain often see the lowest uptake.
This project builds a screening tool that identifies which neighbourhoods have the worst-performing building envelopes using data municipalities already collect. Aggregated gas consumption by postal code provides the thermal signal, HDD regression extracts the envelope response, and MPAC property data normalizes for building stock differences. The output is a prioritized list of postal codes where retrofit incentive programs are most likely to find uptake and savings — enabling proactive, geographically targeted outreach rather than passive, first-come-first-served allocation.
Notebooks
- 01 Data Acquisition & Exploratory Analysis
Ingests aggregated residential gas consumption by postal code, regional HDD data, and MPAC property tax roll summaries. Performs EDA on gas-temperature relationships across 150 synthetic postal codes.
Pythonpandasnumpymatplotlibseaborn - 02 HDD Regression & Normalization
Runs OLS regression per postal code, applies quality filters, normalizes slopes by MPAC building stock characteristics, and produces a thermal intensity metric for neighbourhood ranking.
Pythonpandasnumpyscipymatplotlib - 03 Neighbourhood Prioritization & Targeting
Ranks postal codes by normalized thermal intensity within structure type groups, identifies priority neighbourhoods for retrofit incentive targeting, and tests sensitivity to normalization assumptions.
Pythonpandasnumpymatplotlibseaborn