Perovskite Properties · foundry-ml
dataset_perovskite_tec
A curated dataset of 137 experimental thermal expansion coefficients for perovskite oxides across temperature ranges.
What dataset_perovskite_tec is
This dataset provides experimental thermal expansion coefficients (TEC) for perovskite oxides, compiled to support machine learning models in materials science, chemistry, and physics. It was developed by Kevin P. McGuinness, Anton O. Oliynyk, Sangjoon Lee, Beatriz Molero-Sanchez, and Paul Kwesi Addo, and published through the Materials Data Facility and Foundry-ML.
Predicting how perovskite oxides expand under thermal changes is essential when designing functional ceramics and energy materials that undergo varying operational temperatures. The dataset supplies standardized experimental measurements alongside compositional and temperature data to train and validate thermal property regression and classification models.
How it was made
The dataset is structured as tabular records containing 137 experimental measurements. Each entry links a specific perovskite material composition, identified with its crystallographic sites, to an initial temperature, a final temperature, and the measured thermal expansion coefficient target.
Key facts
- Data format
- Tabular (Parquet)
- Record count
- 137 entries
- Target property
- Thermal expansion coefficient (TEC)
- Target units
- x10^-6 K^-1
- Input features
- Composition (with sites), Initial Temperature (degC), Final Temperature (degC)
- Task types
- Tabular regression, tabular classification
- Release year
- 2023
- DOI
- 10.18126/avvr-p174
What you can use it for
- Perovskite TEC Regression
- Train supervised machine learning models to predict experimental thermal expansion coefficients from site-specific chemical compositions and defined temperature intervals.
- Thermal Stability Screening
- Screen prospective perovskite oxide compositions for compatible thermal expansion behaviour across specified operating temperature ranges.
- Model Benchmarking
- Evaluate tabular regression algorithms on small-sample experimental materials science datasets.
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/dataset_perovskite_tec at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 137 rows of default/train
Composition
string_text- min
- 22
- median
- 22
- mean
- 22.22
- max
- 24
Final Temperature (degC)
int- min
- 400
- median
- 900
- mean
- 863.66
- max
- 1,000
Initial Temperature (degC)
int- min
- 25
- median
- 30
- mean
- 46.51
- max
- 700
TEC (x10^-6 K^-1)
float- min
- 9.11
- median
- 15.30
- mean
- 16.06
- max
- 29.10
Strengths and limits
Good at
- Ground-truth target values are derived directly from experimental measurements rather than computational approximations.
- Material compositions include crystallographic site designations.
- Includes explicit initial and final temperature boundaries for each thermal expansion measurement.
Watch out for
- The dataset is limited to 137 samples, which restricts training data volume for data-intensive models.
- Records are confined specifically to perovskite oxides and do not span other structural classes.
Questions
Can I look inside dataset_perovskite_tec 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 dataset_perovskite_tec under?
Its authors declare other. Read the terms on the original page before reusing the data.
What units are used for the thermal expansion coefficient?
The target property, TEC, is reported in units of x10^-6 K^-1.
What features are provided as model inputs?
The input features are material composition (including crystallographic sites), initial temperature in degrees Celsius, and final temperature in degrees Celsius.
How many records are included in the dataset?
The dataset contains 137 experimental records organized into a single training split.
How to cite it
McGuinness, Kevin P., Oliynyk, Anton O., Lee, Sangjoon, Molero-Sanchez, Beatriz, Addo, Paul Kwesi. Machine-learning prediction of thermal expansion coefficient for perovskite oxides with experimental validation. Materials Data Facility, 2023. DOI: 10.18126/avvr-p174.
Written from dataset_perovskite_tec'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.