Skip to Content
MDPIMDPI

Advanced

DataData
PDF
  • Editorial
  • Open Access

13 June 2026

3 Pages

Data Science for Chemistry, Energy and Materials: Building the Foundations for Data-Driven Discovery

Institute of Applied Physics and Materials Engineering, University of Macau, Macau, China
Data2026, 11(6), 144;https://doi.org/10.3390/data11060144
This article belongs to the Section Data Science for Chemistry, Energy and Materials
Version Notes
Order Reprints
The emergence of data science has fundamentally transformed the landscape of scientific research. Across chemistry, energy, and materials science, the increasing availability of large-scale experimental and computational data, coupled with advances in machine learning (ML), artificial intelligence (AI), high-performance computing, and automated experimentation, has created unprecedented opportunities for accelerating scientific discovery. Data-driven approaches are increasingly complementing and, in some cases, reshaping the traditional paradigm of theory, simulation, experimental synthesis and characterization. The integration of data science into chemistry, energy, and materials research has given rise to several rapidly growing fields, including materials informatics, cheminformatics, energy informatics, scientific machine learning, autonomous laboratories, and AI-assisted scientific discovery. Simultaneously, advances in scientific databases, digital infrastructures, and FAIR (Findable, Accessible, Interoperable, and Reusable) data principles have established the foundation for a more open, reproducible, and collaborative scientific ecosystem [1].
The rapid growth of scientific databases has played a pivotal role in enabling data-driven discovery. Over the past decade, large-scale initiatives such as the Materials Project [2], Open Quantum Materials Database (OQMD) [3], AFLOW [4], NOMAD [5], and JARVIS [6] have generated millions of computational materials records, creating an unprecedented resource for exploring composition–structure–property relationships. These databases have enabled high-throughput screening of candidate materials for applications ranging from batteries and catalysts to thermoelectrics and photovoltaics. By combining density functional theory (DFT) calculations with automated workflows and machine learning models, researchers can rapidly identify promising materials while significantly reducing computational and experimental costs. In parallel, chemistry has witnessed similar developments through the establishment of molecular databases, reaction repositories, and chemical knowledge graphs. Such resources have facilitated the application of machine learning to molecular property prediction, reaction optimization, catalyst discovery, and drug design. The increasing adoption of FAIR data principles [1] has further accelerated data sharing and interoperability, promoting reproducibility and facilitating cross-disciplinary collaboration.
Machine learning has become one of the most influential tools in modern chemistry and materials science. Early studies focused primarily on property prediction using handcrafted descriptors and statistical learning methods. Today, advanced machine learning architectures can directly learn from atomic structures, molecular graphs, spectra, and images. These advances have significantly expanded the scope of materials informatics and cheminformatics, enabling accelerated screening, inverse design, and optimization across broad classes of chemical systems. The energy sector has emerged as one of the most active application areas for data science. Machine learning methods are increasingly being employed to optimize battery materials, fuel cells, electrocatalysts, solar absorbers, hydrogen production systems, carbon capture technologies, and energy-storage devices [7,8,9].
Perhaps the most transformative development in recent years has been the emergence of foundation models and generative AI for scientific research. Inspired by advances in natural language processing and computer vision, foundation models seek to learn transferable representations from massive datasets that can subsequently support a wide range of downstream scientific tasks.
In chemistry and materials science, large language models (LLMs), graph foundation models, and multimodal AI systems are increasingly being developed to integrate scientific literature, molecular structures, crystal structures, spectroscopy, microscopy, simulations, and experimental observations [10,11,12,13]. Notable examples include AI agents capable of autonomous scientific reasoning [10], chemistry-oriented LLM frameworks such as ChemCrow [11], and large-scale materials discovery platforms such as GNoME [12], which demonstrated the identification of hundreds of thousands of potentially stable inorganic materials.
Another rapidly emerging frontier is autonomous experimentation. The integration of robotics, machine learning, active learning, and automated characterization has enabled the development of self-driving laboratories capable of performing closed-loop scientific discovery [14,15,16,17]. In these systems, experiments are performed automatically, data are analyzed in real time, predictive models are updated continuously, and subsequent experiments are selected autonomously to maximize information gain or optimize targeted objectives.
However, despite remarkable progress, significant challenges remain. High-quality data remain the most critical limiting factor for many data-driven approaches. Scientific datasets are frequently fragmented, incomplete, inconsistent, or poorly annotated. Experimental data often lack sufficient metadata describing measurement conditions, uncertainties, and sample histories.
The development of community-wide standards for data representation, metadata collection, and data validation remains a major priority. Modern scientific investigations generate highly heterogeneous datasets, including spectroscopy, microscopy, diffraction, imaging, simulations, sensor measurements, text, and literature data. Methods capable of integrating these diverse information sources into coherent scientific representations remain in their infancy. Unlike many commercial AI applications, scientific research often operates under severe data limitations. Experimental datasets are frequently small, noisy, and expensive to generate. Methods that improve transferability, uncertainty quantification, and learning efficiency will therefore be critical for future progress.
The “Data Science for Chemistry, Energy and Materials” section seeks to provide a multidisciplinary platform for research at the intersection of data science and the physical sciences. By bringing together researchers from chemistry, materials science, physics, engineering, computer science, and data science, this section aims to accelerate the development of innovative methodologies and foster a more open, collaborative, and reproducible scientific ecosystem.

Funding

This research received no external funding.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Wilkinson, M.D.; Dumontier, M.; Aalbersberg, I.J.; Appleton, G.; Axton, M.; Baak, A.; Blomberg, N.; Boiten, J.W.; da Silva Santos, L.B.; Bourne, P.E.; et al. The FAIR Guiding Principles for Scientific Data Management and Stewardship. Sci. Data 2016, 3, 160018. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  2. Jain, A.; Ong, S.P.; Hautier, G.; Chen, W.; Richards, W.D.; Dacek, S.; Cholia, S.; Gunter, D.; Skinner, D.; Ceder, G.; et al. The Materials Project: A materials genome approach to accelerating materials innovation. APL Mater. 2013, 1, 011002. [Google Scholar] [CrossRef] [Scilit]
  3. Saal, J.E.; Kirklin, S.; Aykol, M.; Meredig, B.; Wolverton, C. Materials design and discovery with high-throughput density functional theory: The open quantum materials database (OQMD). JOM 2013, 65, 1501–1509. [Google Scholar] [CrossRef] [Scilit]
  4. Curtarolo, S.; Setyawan, W.; Hart, G.L.; Jahnatek, M.; Chepulskii, R.V.; Taylor, R.H.; Wang, S.; Xue, J.; Yang, K.; Levy, O.; et al. AFLOW: An automatic framework for high-throughput materials discovery. Comput. Mater. Sci. 2012, 58, 218–226. [Google Scholar] [CrossRef] [Scilit]
  5. Draxl, C.; Scheffler, M. NOMAD: The FAIR concept for big data-driven materials science. MRS Bull. 2018, 43, 676–682. [Google Scholar] [CrossRef] [Scilit]
  6. Choudhary, K.; Garrity, K.F.; Reid, A.C.; DeCost, B.; Biacchi, A.J.; Hight Walker, A.R.; Trautt, Z.; Hattrick-Simpers, J.; Kusne, A.G.; Centrone, A.; et al. The joint automated repository for various integrated simulations (JARVIS) for data-driven materials design. npj Comput. Mater. 2020, 6, 173. [Google Scholar] [CrossRef] [Scilit]
  7. Butler, K.T.; Davies, D.W.; Cartwright, H.; Isayev, O.; Walsh, A. Machine Learning for Molecular and Materials Science. Nature 2018, 559, 547–555. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  8. Schmidt, J.; Marques, M.R.; Botti, S.; Marques, M.A. Recent advances and applications of machine learning in solid-state materials science. npj Comput. Mater. 2019, 5, 83. [Google Scholar] [CrossRef] [Scilit]
  9. Wang, A.Y.T.; Murdock, R.J.; Kauwe, S.K.; Oliynyk, A.O.; Gurlo, A.; Brgoch, J.; Persson, K.A.; Sparks, T.D. Machine learning for materials scientists: An introductory guide toward best practices. Chem. Mater. 2020, 32, 4954–4965. [Google Scholar] [CrossRef] [Scilit]
  10. Boiko, D.A.; MacKnight, R.; Kline, B.; Gomes, G. Autonomous chemical research with large language models. Nature 2023, 624, 570–578. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  11. Bran, A.M.; Cox, S.; Schilter, O.; Baldassari, C.; White, A.D.; Schwaller, P. Augmenting large language models with chemistry tools. Nat. Mach. Intell. 2024, 6, 525–535. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  12. Merchant, A.; Batzner, S.; Schoenholz, S.S.; Aykol, M.; Cheon, G.; Cubuk, E.D. Scaling deep learning for materials discovery. Nature 2023, 624, 80–85. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  13. Choudhary, K.; Wines, D.; Li, K.; Garrity, K.F.; Gupta, V.; Romero, A.H.; Krogel, J.T.; Saritas, K.; Fuhr, A.; Ganesh, P.; et al. JARVIS-Leaderboard: A large scale benchmark of materials design methods. npj Comput. Mater. 2024, 10, 93. [Google Scholar] [CrossRef] [Scilit]
  14. Burger, B.; Maffettone, P.M.; Gusev, V.V.; Aitchison, C.M.; Bai, Y.; Wang, X.; Li, X.; Alston, B.M.; Li, B.; Clowes, R.; et al. A mobile robotic chemist. Nature 2020, 583, 237–241. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  15. MacLeod, B.P.; Parlane, F.G.; Morrissey, T.D.; Häse, F.; Roch, L.M.; Dettelbach, K.E.; Moreira, R.; Yunker, L.P.; Rooney, M.B.; Deeth, J.R.; et al. Self-driving laboratory for accelerated discovery of thin-film materials. Sci. Adv. 2020, 6, eaaz8867. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  16. Häse, F.; Roch, L.; Aspuru-Guzik, A. Next-Generation Experimentation with Self-Driving Laboratories. Trends Chem. 2019, 1, 282–291. [Google Scholar] [CrossRef] [Scilit]
  17. Karniadakis, G.E.; Kevrekidis, I.G.; Lu, L.; Perdikaris, P.; Wang, S.; Yang, L. Physics-informed machine learning. Nat. Rev. Phys. 2021, 3, 422–440. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Development and Multireader Evaluation of Radiological RAG-System
Previous
A Comprehensive Lipidomic Dataset of Melittin-Induced Lipid Metabolism Alterations in Triple-Negative Breast Cancer Cells
Next