PSDI Data Lakehouse
PSDI data lakehouse offers a programmatic access (a machine interface) to datasets that can be queried in the Cloud without the need to download them. The datasets offered currently are:
| Name | Description | Records | References | Terms of use |
|---|---|---|---|---|
| Open Materials 2024 (OMat24) | DFT calculations on inorganic bulk materials; 400 million CPU core-hours | >100M | arXiv:2410.12771 and Hugging Face OMat24 | As in Hugging Face OMat24 |
| Open Molecules 2025 (OMol25 parsed version) | DFT calculations on small molecules, biomolecules, metal complexes, and electrolytes; 6 billion CPU core-hours | >100M | arXiv:2505.08762 and Hugging Face OMol25 | As in Hugging Face OMol25 |
| Materials Project (MP) – AWS Open Data | Physical properties derived from DFT calculations (17 datasets) | ~200k compounds | doi.org/10.1038/s41563-025-02272-0 | As in Materials Project Terms of Use |
| Materials Project – Contribs (MPtrj & MatPes) | DFT datasets for training machine-learning interatomic potentials | >2M | doi.org/10.1038/s42256-023-00716-3 | As in Materials Project Terms of Use |
Acknowledgements and disclaimers
We do not claim any rights on top of the terms of use of the original datasets. We converted the original datasets in their lakehouse respresentations with a reasonable attention to quality, but without giving any warranty for the completeness or integrity of data transformation, or for suitability of the resulted format to any particular needs.
If you use the datasets in PSDI lakehouse, we suggest you respect the terms of use of the original datasets, including giving proper attribution and due credit to data originators. If you want giving an additional credit to PSDI lakehouse, you can mention it as: PSDI Data Lakehouse. https://lakehouse.psdi.ac.uk/
Access to the lakehouse
Lakehouse is a multi-layered application, with users having access to its consumption layer. The consumption layer can be accessed with SQL requests or with dataframes requests in Python.
Jupyter and Marimo notebooks in Notebooks examples contain links to the actual consumption layer and are a good starting point to explore the lakehouse content and functionality.
When you access the consumption layer, you are prompted to login with the PSDI user account or create a new one:
More information
More information about the PSDI lakehouse can be obtained from a presentation about it at http://purl.org/net/epubs/work/66187592 and a technical report at https://doi.org/10.5286/stfctr.2026011
Supplying feedback
Feedback on the lakehouse can be supplied at support@psdi.ac.uk
- Creator: Vasily Bunakov
- Last modified date: 2026-07-27
- Citation: Please cite: PSDI Data Lakehouse, https://guidance.psdi.ac.uk/docusaurus-pages/docs/guidance/psdi-resources/data-lakehouse/, PSDI (modified 2026-07-27)
If you would like to contribute content to the PSDI Knowledge Base or have feedback you would like to give on this guidance, please contact us.