Superconductivity · foundry-ml
superconductivity_v1-1
A tabular dataset of experimentally measured superconducting critical temperatures for 16,414 material compositions.
What superconductivity_v1-1 is
superconductivity_v1-1 is a curated materials science dataset compiled by Valentin Stanev, Corey Oses, A. Gilad Kusne, Efrain Rodriguez, Johnpierre Paglione, Stefano Curtarolo, and Ichiro Takeuchi. Published via Foundry-ML and the Materials Data Facility, it compiles empirical critical temperatures for thousands of inorganic compounds to assist machine learning exploration.
Predicting the critical temperature (Tc) — the point below which electrical resistance vanishes — remains a central challenge in condensed matter physics and energy research. This dataset standardises compositional data and experimental measurements to support computational screening and discovery of superconducting candidates.
How it was made
The dataset is organised in a tabular structure comprising 16,414 material entries. Each entry pairs an input chemical composition string with an experimentally observed superconducting critical temperature recorded in Kelvin.
The dataset is configured with a single training split and formatted for tabular regression and classification workflows using standard data science libraries.
Key facts
- Publisher
- Materials Data Facility / Foundry-ML
- Release year
- 2022
- Dataset size
- 16,414 materials
- Data format
- Parquet tabular data
- Input features
- Material composition (name)
- Target variable
- Superconducting critical temperature (Tc in Kelvin)
- Data splits
- train
What you can use it for
- Critical temperature regression
- Train regression models to predict quantitative superconducting critical temperatures directly from chemical composition formulas.
- Superconductor classification
- Categorise candidate chemical compounds into specific critical temperature brackets or binary transition categories.
- Materials screening
- Prioritise candidate material formulations for experimental synthesis in energy and physics applications.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from foundry-ml/superconductivity_v1-1 at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 16,414 rows of default/train
Tc
float- min
- 0
- median
- 5
- mean
- 17.97
- max
- 143
name
string_text- min
- 1
- median
- 15
- mean
- 15.83
- max
- 42
Strengths and limits
Good at
- Provides experimental critical temperature measurements across 16,414 distinct material compositions.
- Features a clear, minimal tabular schema consisting of chemical formulas and critical temperatures in Kelvin.
- Ready for use in both tabular regression and classification pipelines.
Watch out for
- Provides only a single training split, requiring researchers to generate their own validation and test partitions.
- Lacks crystallographic, structural, and synthesis condition data, limiting inputs to composition strings alone.
- Does not include secondary superconducting parameters such as critical magnetic fields or pressure conditions.
Questions
Can I look inside superconductivity_v1-1 without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is superconductivity_v1-1 under?
Its authors declare cc-by-4.0. Read the terms on the original page before reusing the data.
What variables are included in the dataset?
The dataset contains two variables: 'name', representing the material composition, and 'Tc', denoting the experimentally measured superconducting critical temperature in Kelvin.
How many records are included?
There are 16,414 material entries in the dataset.
Which machine learning tasks does it support?
It is structured for tabular regression to predict numerical critical temperatures and tabular classification based on critical temperature boundaries.
Are there pre-defined evaluation splits?
No, the dataset provides only a single train split, so users must create their own cross-validation or holdout evaluation sets.
How to cite it
Valentin Stanev, Corey Oses, A. Gilad Kusne, Efrain Rodriguez, Johnpierre Paglione, Stefano Curtarolo, Ichiro Takeuchi. Machine learning modeling of superconducting critical temperature. Materials Data Facility (2022). DOI: 10.18126/xlfr-hjrn
Written from superconductivity_v1-1's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.