×D2. Workflows and modelling software: opportunities, best practice and case studies

Most scientific activities can be thought of as workflows which involve sequences of tasks or actions. These tasks often include obtaining input data, processing activities which are then followed by post-processing operations such as visualisation and analysis in support of decision making. Processing steps may involve use of modelling, simulation or analysis programs and may include the outputs of such models being dynamically fed back into the system. Post processing analysis may involve advanced techniques such as Machine Learning to provide insightful analysis which given the data volumes may not be obtainable via traditional methods.

Data archiving, provenance and commercialisation of these workflows as software applications or services are often not considered or are considered late in the development process which creates significant hurdles to maximising the benefit of the work. Very commonly workflows are created as bespoke combinations of software components linked by manual steps or tasks. The presence of manual steps increases the cost of performing the activity and limits its reproducibility and its transfer to others. They are often also conducted on disparate systems, limiting the possibly of automation and integration. Often workflows and the tasks involved are not well documented. This makes packaging capability for commercial use difficult and can limit the range of users to those who have developed the process.

The use of workflow engines allows scientific workflow to be systematised and automating leading to reduction in cost of creation and use, reducing dependence on specific people and skills and supports improved reliability, reproducibility, provenance and commercialisation.

This session provides a forum for researchers to explore the opportunities for improvement of their scientific workflow development, to discuss what is modelling and modelling software best practice and to exchange their experiences in using workflow concepts to improve their scientific research and its impact.

Key topics: Workflows, Best practice, Workspace, Commercialisation