Ecoinformatics Conference Service, International Conference on Ecological Informatics 6

Using Scientific Workflows in Kepler for Conservation Science

Kevin L S Drury, Mathew Jones, Mark Shildhauer, Judith Kruger, James Regetz, Benjamin Leinfelder

Last modified: 2008-09-13

Abstract


The recurring challenges confronting conservation scientists define basic requirements of useful conservation-analysis software. We use conservation analyses designed for monitoring conservation issues in Kruger National Park, South Africa, to demonstrate how Kepler, a system for designing and running scientific workflows, allows conservation scientists to simultaneously satisfy these requirements, within a single framework. Historically, a wide variety of software has been used to satisfy data management and analysis requirements, typically with strengths in these areas, but often with pronounced deficiencies in auditability (i.e., the ability to verify that the analysis does what it purports to do) and transparency (i.e., the ability for others to understand the analysis by examining the software). For example, other software, such as programs for generating diagrams that summarize the analytical approach, is often necessary - even for communicating to specialist audiences, such as those reached by peer-reviewed journals. Kepler addresses these issues by seamlessly integrating the representation of the logical flow of data and analysis with the flow itself. That is, scientific workflows in Kepler communicate diagrammatically, but are executable in the sense that the workflow is bound directly to underlying analytic code. Thus, Kepler combines the strengths of rigorous analysis with transparent communication necessary for replication and for reaching the broad spectrum of audiences comprising the stakeholders of conservation efforts.