An Open-Source Data Science Platform for the Electrochemical Community

Tuesday, 3 October 2017: 16:00
National Harbor 4 (Gaylord National Resort and Convention Center)
R. Masse (University of Washington)
It is no exaggeration to say that the lithium-ion battery has revolutionized the world, and the importance of electrochemistry has been brought to the fore as we look to improve batteries for electric vehicles, biomedical devices, and other critical applications.

However, electrochemical studies – particularly on batteries – are conducted with software tools that were not designed for the scale and the importance of the energy storage industry today. The software provided by hardware vendors often does not readily provide the answers we need, or even allow us to ask the right questions. Legacy software tools are often no better and suffer from severe stability issues. In industry, time is critical and most data goes unanalyzed, putting manufacturers at risk. Other firms build in-house tools that constantly reinvent the wheel.

We believe the community would be better served by working together with tools commensurate with the tasks we face. In recent years, several open-source software and hardware tools have been developed by the electrochemical community. However, most have only garnered modest traction. We believe this is driven by the high barrier of adoption, since most electrochemists are not also software developers, and because a significant amount of time and resources must be dedicated to supporting such projects.

Here we present an open-source web-based tool that lays the foundation for collaboration, and addresses the issues described above. Our platform is designed to accomplish the following:

1) Automate rote data analysis. Hours of analysis can be done in minutes or seconds to generate standard plots and statistics for cell capacity, voltage profiles, and cyclic voltammetry, as well as more involved analyses like incremental capacity or other user-defined functions. This is particularly important for batteries where both the number of electrochemical cells and cycles per cell may be very large and cumbersome to analyze thoroughly.

2) Reduce the barrier to enter and contribute to the platform. We provide APIs for a common framework for others to build upon. This saves time compared to developing a one-off application from scratch, and allows for greater impact when packaged together with a suite of other tools.

3) Catalyze electrochemical data science. The growing field of the data science allows for new tools to be developed to ask new questions and glean previously invisible insights from battery performance data.

We believe this software platform provides the missing infrastructure required to meet the demands of the electrochemical community today, and rise to the challenges of tomorrow.