Ionic Conductivity Data for Li and Na Solid-State Electrolytes with Machine Learning Code

A repository containing curated datasets of room-temperature ionic conductivities for lithium-based and sodium-based solid-state electrolytes. The materials include CSV files for Li, Na, and combined datasets, a Jupyter notebook for regression modeling, and supplementary Excel tables and figures. All data was obtained through a combination of large language model-assisted extraction and manual curation to support the associated research article.

Use Cases

• Training regression models to predict ionic conductivity based on material composition.
• Benchmarking machine learning models for materials property prediction using curated datasets.
• Reproducing the classification and regression analyses described in the original research article.
• Featurizing host materials for solid electrolytes using one-hot encoding of mobile ions.

Strengths

• Includes three distinct curated datasets (Li, Na, and combined) specifically for solid-state electrolytes.
• Provides full reproducibility with a 175.6 MB repository containing the original Jupyter notebook code and supplementary information.
• Data collection method combines large language model-assisted extraction with manual curation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale modeling.
• The description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Author Dane Morgan, via figshare.

Collection Method

Combination of large language model-assisted extraction and manual curation.

Time Range

Room-temperature measurements; specific temporal coverage is not stated.

Freshness

Last updated 2026-05-06 08:10:13; freshness should be verified.

Geography

Spatial coverage is not specified.