Tools

Energy Toolkit provides two tools for energy forecasting. Both follow the same design philosophy (a Python generator, a data API, and an interactive web frontend) but address different questions about the energy system.

⚡

Demand

How much electricity will we need? Scenario-based demand projections from historical data, with interactive charts and geographic maps.
🌱

Generation

What do we need to build? PyPSA-based optimization to plan renewable capacity (solar, wind, storage) with weather-informed modeling.

Shared design philosophy

Both tools are built around the same three-layer architecture. This is a deliberate design choice that makes each tool independently forkable, deployable, and customizable while sharing a common mental model.

Generator

A Python engine that takes input data and configuration, runs a model, and produces structured output. The generator is the only part that needs domain expertise.

  • Demand: Historical data + transformations → Parquet
  • Generation: Weather + demand + PyPSA → capacity plans

API

A data layer that serves generator output through standardized endpoints. Static JSON for metadata, dynamic queries for time series and scenario data.

  • Demand: OpenAPI 3.1 + DuckDB on Parquet
  • Generation: File-based API (local or HTTP server)

Frontend

An interactive web application for exploring the data. Charts, maps, scenario comparison, designed for multiple audiences from the public to domain experts.

  • Demand: SvelteKit + LayerChart + Mapbox GL
  • Generation: Streamlit + Plotly + Folium

Designed to be forked

The tools are not SaaS products. They are starting points. Fork a repo, swap in your own data and models, and deploy. The architecture is designed to make this straightforward.

Flexible inputs

  • Any region: Provide your own GeoJSON boundaries and demand data. The Demand tool ships with Swedish data; adapt it for your country or region.
  • Any scenario model: The generators are modular. Replace the transformation functions with your own forecasting models while keeping the data pipeline intact.
  • Configuration-driven: Scenarios, geographies, segments, and parameters are defined in config files (YAML or JSON). Change the config, regenerate, and the entire stack updates.

Flexible deployment

  • Run locally: Both tools work on a laptop with no cloud services. Good for development and internal use.
  • Deploy to the cloud: Both tools run on any standard cloud platform. Bring your own infrastructure; the tools have no vendor lock-in.
  • Static or dynamic: Frontends generate fully static sites that can be served by any static host or CDN. APIs can run as containers or serve pre-built files.

Maturity

The two tools are at different stages of development. Both are functional and deployed, but the Demand tool has a more mature framework architecture.

⚡

Demand

  • ✓ Full-stack framework with clear separation
  • ✓ Comprehensive test coverage (60+ tests)
  • ✓ CI/CD pipeline with GitHub Actions
  • ✓ Documentation, coding standards, debugging guides
  • ✓ Multi-language support (Paraglide i18n)
🌱

Generation

  • ✓ Working generator with PyPSA optimization
  • ✓ Streamlit dashboard with scenario controls
  • ✓ Flexible API (local filesystem or HTTP server)
  • ✓ Docker deployment support
  • ○ Planned: SvelteKit frontend migration
  • ○ Planned: Enhanced documentation
View live projects