CarbonGraph connects you to a growing set of Life Cycle Assessment (LCA) datasets, sourced from authoritative external databases such as Ecoinvent, IDEMAT, and the NREL USLCI.

In CarbonGraph, a dataset is a versioned collection of LCA processes imported from a larger source database, then adapted so it can be used consistently inside your models.

We emphasize the use of primary data—data you measure, generate, or collect yourself—because it’s the best way to represent real operating conditions. To support complete modeling when primary data isn’t available, CarbonGraph also provides an expanding library of curated secondary datasets.

All models and datasets created within your CarbonGraph account are version-controlled and can be synced or updated automatically. This allows you to link your proprietary data with reference datasets and maintain organization-specific models that evolve over time.

Don’t see the dataset you need? We’re continually expanding coverage. Reach out to us at support@carbongraph.io to request additional database integrations.

Any model you build in CarbonGraph becomes part of your private dataset. These models are reusable within your organization and can be imported into other assessments as needed—supporting team workflows, repeatability, and internal consistency.

• Private and secure: By default, models are only visible to your team. Sharing is optional and controlled.

• Version-controlled: Track changes, review previous states, and sync updates across linked assessments.

• Tailored to your data: Use your own inputs, operating conditions, and assumptions to reflect real-world performance.

• Integrates with reference datasets: Combine proprietary models with secondary data when needed to fill gaps.

Ecoinvent is one of the most widely used and respected LCI databases globally. It supports consistent, transparent modeling of supply chains across numerous sectors.

Current Versions Available:

Ecoinvent 3.11 (latest stable release as of Nov. 2024)

• Allocation, Cut-off by Classification

• Allocation, Cut-off, EN15804

• Allocation at Point of Substitution*

• Substitution Consequential Long-term*

Ecoinvent 3.9.1

• Allocation, Cut-off by Classification

• Allocation, Cut-off, EN15804*

• Allocation at Point of Substitution*

• Substitution Consequential Long-term*

Note: * Indicates Available Upon Request

Website: https://ecoinvent.org

Strengths:

• Comprehensive sector coverage with over 18,000 datasets covering energy, chemicals, materials, agriculture, transport, and more.

• Transparent documentation of system boundaries, allocations, and data sources.

• Regular updates reflecting scientific research, industry changes, and evolving standards.

• Multiple system models (including Cut-off, Allocation at the Point of Substitution, and Consequential modeling).

Coverage and Characterization:

• Global and regional datasets, including detailed upstream processes and regional electricity mixes.

• Compatible with a range of LCIA methods including ReCiPe 2016, TRACI 2.1, EF 3.0/3.1, IMPACT World+, and EPS 2020.

The USLCI dataset, developed by the National Renewable Energy Laboratory (NREL), is a publicly available database representing U.S.-specific processes. It is part of the Federal LCA Commons and provides a valuable foundation for regionally relevant LCAs.

Current Version: v1.2025-03.0 (Elementary Flow List v1.3.0)

Website: https://www.lcacommons.gov

Strengths:

• U.S.-specific data reflecting domestic technologies, infrastructure, and energy mixes.

• Transparent documentation and open access.

• Flexible modeling options through multiple dataset variants.

Coverage and Characterization:

• Approximately 1,500 processes spanning fuels, electricity, transportation, construction materials, and waste treatment.

• Supports LCIA methods such as ReCiPe 2016 (midpoint and endpoint) and TRACI 2.1.

Dataset Variants Available:

• Gate-to-Gate: Original unit processes as published by NREL, representing individual process steps without upstream or downstream linkages.

• Cradle-to-Gate: Pre-linked system models that incorporate upstream processes, supporting ready-to-use modeling from resource extraction to the process gate.

IDEMAT, developed by TU Delft, is well-suited for early-stage product design and material selection. It emphasizes engineering use cases and eco-cost modeling in support of circular design.

Current Version: IDEMAT 2023 – Cut-off by Classification

Website: https://www.ecocostsvalue.com

Strengths:

• Strong alignment with product design tools and academic eco-design methodologies.

• Emphasizes transparency and straightforward integration into engineering workflows.

Coverage and Characterization:

• Over 2,000 processes covering common materials, plastics, metals, and wood products.

• Strong focus on European infrastructure and energy systems.

• Supports ReCiPe 2016, Eco-costs, CED, and both midpoint and endpoint methods.

Selecting the appropriate dataset is a key part of LCA best practices, as described in ISO 14044. The right dataset choice improves relevance, transparency, and alignment with your assessment goals.

Consider the following factors:

• Geographic scope: Use regional datasets when modeling region-specific systems—especially electricity, transport, and resource extraction.

• Sector specificity: Some datasets have stronger coverage in particular industries such as chemicals, agriculture, or construction.

• Data quality and completeness: Look for transparency in assumptions, sources, and methodological choices.

• LCIA compatibility: Ensure the dataset is compatible with the impact assessment methods you plan to use (e.g., ReCiPe, TRACI).

• Update frequency and maintenance: Prefer datasets that are actively maintained and regularly updated.

CarbonGraph enables you to combine multiple datasets within a single model. This flexibility lets you use the best available data for each part of your system while maintaining overall consistency.

Each dataset available in CarbonGraph reflects the methodological choices and data structure of its original source. We recommend reviewing the official documentation provided on each dataset’s website for full details regarding data quality, assumptions, and conformance to standards such as ISO 14040 and 14044.

While CarbonGraph performs validation and quality checks during dataset ingestion, harmonizing diverse datasets into a consistent modeling environment can be challenging. Differences in system boundaries, flow naming, or characterization factors can introduce inconsistencies—and despite our best efforts, errors may occasionally persist.

If you identify an issue or suspect a discrepancy, please contact us at support@carbongraph.io. We actively investigate all reports and appreciate your help maintaining the quality and reliability of shared data resources.