Webinar: Data Revival

This webinar continues our look at our pathfinder activities in the current phase. The subject of this webinar is extraction of knowledge from lab books, which is being undertaken as part of our Pathfinder 2 & 3 activities in conjunction with spin out company Data Revival. This webinar will be presented by Samuel Munday from Data Revival.

The recording of this webinar is now available on Youtube

Making the intangible tangible: The journey from lab notebook to digital insight

The University of Southampton’s chemistry department has accumulated a wealth of chemical knowledge over the years, some of which has been captured in over 2000 lab books. The vast majority of these have been sat gathering dust in a cupboard for a long time, unable to be destroyed due to both the value of knowledge they hold and their importance for Health and Safety reporting. However, this knowledge is intangible and difficult to access, offering no value to the department whilst taking up space and presenting a fire risk. We present our work on turning this unstructured resource into a structured accessible database that holds FAIR data open to analytics and intelligent search. Our system utilises AI-driven natural language processing techniques, as well as chemical structure recognition, to extract all the types of chemical information required to create a useful, searchable database. This database unlocks 3000 chemist years of knowledge and enables more efficient and accurate future research. In this webinar we discuss the process of digitising such an archive effectively, the AI tools we have created to work with such unstructured knowledge at scale, the utility of the digital database created for the chemistry department, and the feedback received from the department on the system’s potential for further development.

Biography

Samuel Munday is a senior research assistant at the University of Southampton as well as being the co-director of the fledgling spin out Data Revival. He first became interested in scientific data management whilst building a predictive analytics platform for polymeric materials and realising that a lot of key data resided in a form incomprehensible to computers. This has led to the development of a series of tools for unstructured chemical data extraction and structuring, mainly used for turning hand written lab notebooks into structured searchable databases at scale. He is currently leading the development of this platform which is beginning to show signs of success with both academic researchers and commercial partners.

Watch the recording

You can watch the recording of this webinar via our Youtube channel. Slides are available on zenodo.