FORCE Documentation

FORCE: French Open Research Catalogue of Energy.

A high-performance Python ETL pipeline for French power-system research, targeting reproducible multi-source data acquisition (RTE, Enedis, Météo-France, ODRÉ), multi-frequency transformation, and ML-based imputation.

Getting Started

• FORCE: French Open Research Catalogue of Energy - A High-Performance Python Pipeline
  • Abstract
  • Documentation & Scientific Methodology
  • Installation & Usage
  • Scientific Method: Data Harmonization
  • Dataset Output Structure
  • Project Structure
  • Dataset Access
  • Citation
  • License & Copyright

Data Sources

• RTE Power Data Pipeline
  • Part 1: Collection Module (collect_rte.py)
  • Part 2: Transformation Module (transform_rte.py)
  • Final Feature Schema
• Enedis Distribution Grid Pipeline
  • Part 1: Collection & Ingestion Modules
  • Part 2: Pre-Processing Module (enedis_utils.py)
  • Part 3: Transformation Module (transform_enedis.py)
  • Final Feature Schema
• Meteorological Data Pipeline
  • Part 1: Collection Module (collect_meteo.py)
  • Part 2: Transformation Module (transform_meteo.py)
  • Part 3: National Aggregation Module (transform_meteo_national.py)
• Solar Radiation & Climate Projection Pipeline
  • Nomenclature and Physical Quantities
  • Part 1: PVGIS (Historical Baseline, Primary Source)
  • Part 2: ERA5 (Complementary Reanalysis)
  • Part 3: Climate Projection (Delta-Change Method)
  • Part 4: Source Comparison and Audit
  • Data Coverage and Requirements
• Production Assets Pipeline (ODRÉ)
  • Part 1: Collection Module (collect_production_assets.py)
  • Part 2: Spatial Transformation Logic
• Calendar and Temporal Feature Pipeline
  • Part 1: Collection Module (collect_calendar.py)
  • Part 2: Transformation Module (transform_calendar.py)
  • Final Feature Schema
  • Output Integration

Methodology

• Machine Learning Imputation Strategy
  • Objective
  • Imputation Engines: TS-ICL (Default) vs. XGBoost
  • The “Cascade” Algorithm (XGBoost Engine)
  • Limitations & Stationarity Assumptions
  • Machine Learning Core: Engine Details (model.py & predictive.py)
• Data Hierarchies & Algebraic Relationships
  • 1. Energy Balances: Transport vs. Distribution
  • 2. Load & Production Decompositions
  • 3. RTE National vs. the 12 Metropolitan Regions (Approximate)
  • 4. RTE Family Totals with Unlisted Sub-streams (Partial)
  • 5. RTE Golden Record (Maturity Coalesce)
  • Summary of Relationships
• TAM Feature Extraction
  • Objective
  • Running the Extraction
  • Outputs
  • Feature Engineering
  • Verification

Community

• Contributing to FORCE
• Contributor Covenant Code of Conduct