Promethean

Summary

Promethean is a real-time observatory for the global energy transition, integrating capacity, generation, emissions, and finance data across 200+ economies. The platform unifies 49 async collectors from IEA, Ember, OWID, IRENA, EIA, EDGAR, Climate Watch, and other authoritative sources into a fuel and sector aware warehouse. Four analytical engines turn raw observations into transition pathways, decarbonization gaps, and capital flow diagnostics.

What it is

Promethean is a production data platform that tracks the worldwide shift from fossil to low-carbon energy with the same rigor a credit risk team would apply to a sovereign book. The system runs 49 asynchronous collectors against primary sources, IEA, Ember, OWID, IRENA, EIA, EDGAR, Climate Watch, World Bank, WRI, and Global Energy Monitor, then normalizes every observation into a fuel and sector aware schema. Annual data is stored with a month=0 sentinel, sub-annual data with months 1 through 12, and ISO3 country codes are enforced upstream of every query. The SQLite warehouse runs in WAL mode with five tables for sources, series, observations, countries, and a collection log that supports operator forensics.

Four engines sit on top of the warehouse. The capacity engine reconciles installed renewable capacity, additions, and retirements by technology. The emissions engine consolidates GHG inventories and computes carbon intensity by economy and sector. The transition engine produces electricity mix decompositions, fossil displacement curves, and policy alignment indicators. The finance engine tracks climate and energy investment flows, NDC commitments, and subsidy patterns. A FastAPI service exposes read-only data routes for series, summaries, persistence statistics, cross-country comparators, and panel extracts, alongside engine-specific analytics for capacity, emissions, transition, and finance. APScheduler orchestrates staggered UTC collection, and an admin API surfaces collection-log diagnostics for operators.

Methodology

• Capacity engine reconciles installed capacity, gross additions, and retirements by technology and country, sourced from IRENA and EIA panels.
• Emissions engine consolidates GHG inventories from EDGAR, Climate Watch, and OWID, then derives carbon intensity per unit of GDP and per unit of generation.
• Transition engine decomposes electricity mix shifts using Ember monthly data, isolating renewable share gains, fossil displacement, and demand growth contributions.
• Finance engine tracks climate and energy investment flows, NDC commitments, and fossil subsidy estimates from Climate Watch and World Bank series.
• Fuel and sector aware schema stores observations with explicit fuel and sector tags, enabling clean roll-ups across coal, gas, oil, nuclear, hydro, solar, wind, and bioenergy.
• Annual data uses a month=0 sentinel, sub-annual data uses months 1 through 12, and ISO3 country codes are normalized to uppercase before every query.
• Per-series analytics include AR(1) persistence, half-life estimation, YoY growth, volatility, and shock detection, exposed under /api/data/series/{code}.
• Cross-country comparators rank economies on any indicator and feed the panel extract endpoint for long-format multi-series research.

Data sources

• IEA (EV Outlook, PVPS programme)
• Ember (electricity, monthly generation, capacity factor)
• Our World in Data (energy, CO2, electricity mix, renewables)
• IRENA (capacity statistics, renewable cost database)
• EIA international energy statistics
• EDGAR global GHG inventory
• Climate Watch (GHG, NDC tracker)
• World Bank (renewable energy, electricity sources, GHG, air quality), WRI PowerWatch, Global Energy Monitor

Deliverables when used in engagements

• Read-only data API covering series, summaries, persistence, comparators, and panel extracts
• Capacity, emissions, transition, and finance engine endpoints with country and fuel breakouts
• Operator dashboard with collection-log diagnostics and source health
• Country profile briefs assembling all series for a single economy
• Reproducible CLI for init-db, collect, collect-all, and status reporting