An interactive tool that compares emission factors across multiple databases and countries, making methodology disagreements and their causes visible.
When calculating emissions from a business activity, multiple public databases publish emission factors that frequently disagree. This tool normalizes factors to a common unit (kg CO2e), calculates variance, and diagnoses why they differ: real physical differences, methodological choices, or data provenance issues.
- By Source: Compares the same activity across EPA, DEFRA, and GHG Protocol (8 activities with 2+ sources, 20 total)
- By Country: Compares the same activity across the top 5 emitting countries (US, China, India, Germany, Russia) for 4 key activities
- Geography dominates. Most variance comes from comparing US factors to UK factors, not from methodology disagreement.
- Source independence is an illusion. GHG Protocol compiles from IPCC, DEFRA, and EPA. Apparent "three-source agreement" may actually be circular citation.
- Per-passenger metrics amplify occupancy. Indian buses emit 15 g CO2/pkm vs US at 110 g, a 7x gap driven by ridership, not technology.
- Some factors are genuinely universal. Natural gas shows 0% cross-country variance because combustion chemistry doesn't vary by country.
| Source | Edition | Focus |
|---|---|---|
| EPA GHG Emission Factors Hub | January 2025 | US-focused |
| UK DEFRA/DESNZ Conversion Factors | June 2025 | UK-focused |
| GHG Protocol Cross-Sector Tools V2.0 | May 2024 | Compiles from IPCC, EPA, DEFRA |
| Ember Climate Global Electricity Review | 2025 | Country-level grid intensity |
| ICCT, UBA TREMOD, IPCC | Various | Country-level transport factors |
Provenance note: GHG Protocol does not publish independent emission factors. It compiles from IPCC, DEFRA, IEA, and EPA. This is itself a finding worth discussing.
app/ # Next.js frontend (static export for GitHub Pages)
data/
raw/ # Original Excel files from each source (unmodified)
normalized/ # Extracted and unit-normalized JSON
country_specific/ # Country-level emission factors for top 5 emitters
scripts/
compare.py # Source-vs-source comparison engine
compare_countries.py # Country-vs-country comparison engine
normalize_*.py # Per-source extraction and normalization
docs/
friction-log.md # Problems encountered during data extraction
data-sources.md # Detailed provenance for each source
- Countries: Limited to top 5 emitters (US, China, India, Germany, Russia)
- Activities: Country comparisons cover 4 high-impact categories (bus, electricity, passenger car, natural gas)
- Data: Uses publicly available sources only. IEA data (paid) would expand coverage significantly.
cd app
npm install
npm run devLearning artifact, not a production tool.