Speaker
Description
Developing new materials requires extensive experimentation in synthesis and characterization, generating vast data sets. To keep this wealth of knowledge and adhere to the FAIR principles, effective data management is essential, involving standardized metadata schemas and integrated analysis tools. NOMAD has recently incorporated structured metadata schemas to manage experimental data from myriad sources, providing a user-friendly way of digitalizing experimental research.
In this contribution, we present a NOMAD analysis plugin, which establishes an integrated analysis workflow along with tools for automation. The plugin allows users to select and link structured data available in NOMAD from their experiments to perform analysis steps. Based on the data, specific standardized analyses may run automatically. The data entry associated with the analysis saves the settings along the output, making the analysis FAIR and thus promoting reproducibility. Additionally, it offers an open Python coding environment via Jupyter Notebooks for custom analysis adjustments.
The NOMAD analysis plugin leverages the platform's sharability and scalability while significantly enhancing its utility by enabling the training and integration of ML models. We illustrate its ML application with a case study on materials synthesis via metalorganic vapor-phase epitaxy. By enabling automated analysis, ensuring reproducibility, and supporting ML applications, the plugin motivates researchers to digitalize their workflows, effectively reducing adoption barriers.