Blended Cement Concrete Database

This repository provides a curated, literature-derived database of blended cement concrete for machine learning, materials informatics, and data-driven research on cementitious materials. The current release contains 8,979 experimental records from 260 publications.

The database includes:

• Mixture proportions: cement, blast-furnace slag, fly ash, silica fume, limestone powder, calcined clay, water, superplasticizer, coarse aggregate, and fine aggregate;
• Curing conditions: age, temperature, and relative humidity;
• Specimen dimensions: volume and aspect ratio;
• Mixture properties: compressive strength; and
• Source information: DOI of the original publication.

The database was constructed through automated information extraction from the scientific literature followed by multiple stages of data cleaning and validation. More than 7,500 records (>80% of the final database) were manually verified. Missing values are marked as null.

Details of the database construction, extraction pipeline, validation procedures, and evaluation are reported in the associated publication, “Large language model-enabled automated data extraction for concrete materials informatics” (https://doi.org/10.1038/s41524-026-02304-6). The human-annotated ground-truth benchmark used to evaluate the extraction pipeline is available as the Concrete Materials Data Extraction Benchmark at https://doi.org/10.5281/zenodo.22132836.

Version note

[2026-08-27] This initial release focuses on mixture proportions, curing conditions, specimen dimensions, and compressive strength. Binder-property variables for the individual cementitious materials—including oxide composition, loss on ignition, specific gravity, and Blaine fineness—are planned for inclusion in a future version of this repository.

Citation

If you use this database, please cite both the associated publication and the dataset:

• Li, Z., Yang, K., He, Q., & Gong, K. (2026). Large language model-enabled automated data extraction for concrete materials informatics. npj Computational Materials. https://doi.org/10.1038/s41524-026-02304-6
• Li, Z., Yang, K., He, Q., & Gong, K. (2026). Blended cement concrete database [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.18080350