Skip to content

technologydata Header Logo

PyPI version License: MIT Python OpenSSF Scorecard pre-commit.ci status

technologydata: techno-economic assumptions for energy models

technologydata is a Python package that supports the management of techno-economic assumptions for energy system models. It provides a structured way to store, retrieve, and manipulate data related to various technologies used in energy systems, including unit-ful parameters, currency conversions, inflation adjustment, and temporal modelling.

The goal of this package is to make energy system modelling easier and more efficient, automating common tasks and transformations to reduce errors and allowing for easier data exchange between models.

Techno-economic catalogues are published in different currencies, price years, physical units, heating-value conventions and parameter naming schemes. Harmonising to combine multiple sources into one energy system model therefore requires a sequence of conversions that is easy to get wrong and rarely recorded.

This package contains a data schema and represents the techno-economic assumptions as Python objects. The objects called Parameter carry not only values ("magnitude"), but also unit information, currency year, energy carrier, heating-value basis, provenance and bibliographic sources. Commonly used conversions are available, parameters can be checked for consistency. Sources stay attached through conversions and arithmetic, and parameters derived with the built-in equations record the formula and input values used.

The package is meant for energy system modellers who:

  • combine assumptions from more than one source, or from a source and their own numbers;
  • convert between currencies, price years, units and heating values without losing track of what was converted;
  • need to know where each number came from and how it was calculated;
  • derive related parameters or check them for consistency, and project values to years their sources do not cover.

Installation

pip install technologydata

Or, using uv:

# Standalone
uv pip install technologydata

# As part of a project
uv add install technologydata

The package requires Python 3.12 or newer. The bundled datasets are installed with it; no additional download step is needed.

New to technologydata? Start with the Overview.

Features of the package

  • Unit-ful parameters. Parameters come with units, including currencies an currency years, heating values and support for (energy) carriers. Unit conversion is handled automatically and unit-compatability is checked for all operations via pint.
  • Currency and inflation adjustment. to_currency() combines exchange-rate conversion and deflation between price years using official inflation exchange rate data from the World Bank or alternatively the International Monetary Fund via pydeflate.
  • Heating-value handling. Parameters can record an LHV or HHV basis and be converted between them with change_heating_value(), using the speicifc energy content depending on the energy carrier.
  • Arithmetic with compatibility checks. Adding parameters with incompatible units, carriers or heating-value bases raises ValueError rather than returning a misleading value.
  • Provenance and source tracking. Every Parameter carries a provenance and source information. Operations and modifications on Parameter objects automatically track provenance information for improved reproducibility and insights.
  • Filtering and export. TechnologyCollection.get() filters by case-insensitive regular expression on name, region, year, case and detailed technology; collections export to pandas.DataFrame, CSV and JSON, and reload from JSON.
  • Interpolation and extrapolation. GrowthModels allow for easy gap filling when data is missing for particular years. Use model fitting and projections with commonly used growth models, including linear, exponential and logistic growth.

Bundled datasets (Batteries included)

The package already ships with some parsed data catalogues for immediate use. Prominent examples are:

  • Energy Storage Catalogue from the Danish Energy Agency (automatically extracted)
  • Annual Technology Baseline 2024 from NREL (manually collected in the old technologydata repository)

The raw files and parsing logic is also shipped alongside to provide the opportunity for verification.

Tip

More packaged data to come! Your contributions are also very welcome!

Example

The Danish Energy Agency storage catalogue provides investment cost of utility-scale lithium-ion battery storage. NREL's ATB 2024 provides similar technology costs, but their currency units do not align. The follow examples loads both datasets and harmonises them to the same units for comparison:

import technologydata

# Load data provided by the DEA
dea = technologydata.DataAccessor(data_source="dea_energy_storage", version="v10").load()
# Load data from NREL's ATB2024
atb = technologydata.DataAccessor(data_source="legacy_input_data", version="v0.13.4").load()

# We get the data specific to battery storage and adjust the currency year
# from 2022 to 2023. By default the inflation adjustment uses World Bank data
us = atb.technologies.get(
    name="battery storage", case="Moderate - Market",
    region=None, year=None, detailed_technology=None,
).to_currency("USD_2023")

# The DEA data is usually valid for the "EU" and the entries carry the region "EU"
# For currency conversion, this is an valid ISO 3166 alpha-3 code that can be used
# for currency conversion and inflation adjustment, so we need to specify a reference country
# like Germany explicitly
eu = dea.technologies.get(
    name="lithium ion battery", case="control",
    region=None, year=None, detailed_technology=None,
).to_currency("USD_2023", overwrite_country="DEU")

for t in eu.technologies:
    if t.year == 2030:
        p = t.parameters["specific investment"].to("USD_2023/kWh")
        print(f"DEA v10  {t.year}: {p.magnitude:7.2f} {p.units}")

for t in us.technologies:
    if t.year == 2030:
        p = t.parameters["investment"]
        print(f"NREL ATB {t.year}: {p.magnitude:7.2f} {p.units}")

DEA v10  2030:  351.74 USD_2023 / kilowatt_hour
NREL ATB 2030:  264.23 USD_2023 / kilowatt_hour

Each converted Parameter retains the sources of the value it came from, so the origin of a number remains traceable after conversion. The currency conversion uses exchange-rate and deflator series retrieved through pydeflate. Data from the World Bank is used in the example above and is downloaded on first use automatically; the data is cached locally and later calls are much faster.

Citing

If you use technologydata in your research, please cite it as:

technologydata: Data for Energy Systems Models.

The package is available at: https://github.com/open-energy-transition/technology-data/tree/prototype-2.

Authors: Johannes Hampp, Fabrizio Finozzi

License

technologydata is released under the MIT license.

Contacts

  • For bugs and feature requests, use the issue tracker.
  • For contributions, open a pull request on GitHub. Ideas, suggestions and problem reports are equally welcome as issues.