Thermal Properties · foundry-ml
dataset_debyet_aflow
A curated dataset of calculated Debye temperatures for 4,896 materials, designed for machine learning property prediction.
What dataset_debyet_aflow is
This dataset provides calculated Debye temperatures for 4,896 materials compiled from the AFLOW repository. The collection was curated by Conrad L. Clement, Steven K. Kauwe, and Taylor D. Sparks as a benchmark resource for machine learning research in materials science and physics.
The Debye temperature is an essential thermal-vibrational descriptor that correlates with crystal lattice dynamics, specific heat, and thermal conductivity. By pairing chemical compositions with corresponding Debye temperatures, the dataset enables computational researchers to build predictive models directly from chemical formulae.
How it was made
The dataset arranges materials information into tabular format with two core fields: chemical composition (formula) as the input feature, and the calculated Debye temperature (target) recorded in Kelvin (K). The records originate from computational data within the AFLOW database.
All 4,896 entries are provided in a single training split, packaged in standard tabular formats such as Parquet for integration into materials informatics workflows.
Key facts
- Entries
- 4,896 materials
- Task
- Tabular regression, tabular classification
- Input feature
- Material composition (formula)
- Target property
- Debye temperature (K)
- Data format
- Parquet, tabular
- Release year
- 2020
- Curators
- Conrad L. Clement, Steven K. Kauwe, Taylor D. Sparks
- Source repository
- AFLOW database via Foundry-ML
What you can use it for
- Debye Temperature Regression
- Train machine learning models to predict the Debye temperature of novel or unmeasured crystalline compounds directly from their chemical formula.
- Materials Informatics Benchmarking
- Benchmark feature engineering schemes, such as composition-based elemental property representations, against a standard materials science thermal property dataset.
- Classification of Thermal Properties
- Categorise inorganic compounds into discrete high- or low-Debye-temperature bins to screen for specific vibrational behaviours.
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_debyet_aflow at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 4,896 rows of default/train
formula
string_text- min
- 2
- median
- 8
- mean
- 7.25
- max
- 15
target
float- min
- 43.87
- median
- 296.94
- mean
- 335.53
- max
- 2,109
Strengths and limits
Good at
- Provides a focused, benchmarked collection of 4,896 materials with standardised thermal property values from the AFLOW database.
- Uses simple composition strings as inputs, making it straightforward to evaluate formula-based featurisation strategies.
- Available in standard Parquet format compatible with standard tabular data tools.
Watch out for
- Does not include explicit three-dimensional crystal structures or crystallographic coordinates, limiting inputs to chemical formulae.
- Comprises only a single predefined train split, requiring researchers to establish their own cross-validation or testing splits.
- Restricted to calculated values derived from the AFLOW database rather than experimental measurements.
Questions
Can I look inside dataset_debyet_aflow 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_debyet_aflow under?
Its authors declare other. Read the terms on the original page before reusing the data.
What properties are included in this dataset?
The dataset contains the material chemical formula as an input feature and the calculated Debye temperature in Kelvin (K) as the target.
How many data points are available?
The dataset contains 4,896 materials entries.
Are crystal structure coordinates provided?
No, the dataset only contains the material composition formula and the target Debye temperature.
What is the origin of the Debye temperature values?
The Debye temperatures are calculated values sourced from the AFLOW materials database.
How to cite it
Clement, Conrad L. and Kauwe, Steven K. and Sparks, Taylor D. Benchmark AFLOW Data Sets for Machine Learning (Debye Temperature). Materials Data Facility, 2020. https://doi.org/10.18126/33r4-8t58
Written from dataset_debyet_aflow'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.