Open Interactive Data Analytics Platform for Chemical-Physics Simulations and Experiments

Advances in computational and experimental research have led to enormous increases in the volume and complexity of scientific data. Research workflows must be significantly enhanced to realize the full potential of scientific discovery. This sentiment is shared by many leaders in the scientific community including Jim Gray, a Turing award winner who stated in his Fourth Paradigm Lecture “We have to do better at producing tools to support the whole research cycle—from data capture and data curation to data analysis and data visualization.” The “reproducibility crisis” highlights this need, scientific publications fail to provide sufficient data, code, and supporting information to reproduce and validate key results—the development of such platforms is of critical importance. The technical approach will leverage best-in-class technologies, and strategically develop new approaches to build an open source platform for chemical-physics simulations, data, and analytics. Features will be developed using a scripting language to glue components together on the server-side, with compiled programs providing optimized implementations and access to existing approaches. The client application will offer a rich, responsive experience using modern web technologies exposed in an electronic notebook interface. The Phase I project developed a prototype software platform with an electronic notebook interface that supported a popular scripting language. It offered search and data analytics, using modern web technologies to deliver interactive visualization of chemical data in the web-based electronic notebook. It scheduled simulations, and made data in the notebook available over the web using web programming interfaces, and in a web page for wider dissemination. The Phase II project will develop a software platform that serves researcher’s needs, offering a data-centric server coupled with an electronic notebook interface, and a modern web application to aid in wider dissemination. It will integrate with major computational facilities to schedule simulations, extract relevant data, and make it available in a form that can be searched, shared, and with reproducibility as a primary driver for chemical-physics simulations. Data representations will be developed and extended, and experimental data used to augment predictive capabilities of in-silico simulations coupled with approaches from machine learning. This project will leverage investments made by the DOE and other agencies to provide a powerful software platform and electronic notebook for the materials design community. It will create a platform that can serve as a reference for the community. The use of permissive open source licensing will make the platform available to all, with a simple deployment strategy offering the ability to keep data close yet make it available to the wider linked semantic web. Instances at national facilities, universities, and the cloud will be made available using Software as a Service (SaaS) models for the cloud instance.