VisTrails: A Python-Based Scientific Workflow and Provenance System (#172)

Description

VisTrails: A Python-Based Scientific Workflow and Provenance System

Presented by David Koop (University of Utah and VisTrails, Inc.)

In this talk, we will give an overview of VisTrails (http://www.vistrails.org), a python-based open- source scientific workflow that transparently captures provenance (i.e., lineage) of both data products and the processes used to derive these products. We will show how VisTrails can be used to streamline data exploration and visualization. Using real examples, we will demonstrate key features of the system, including the ability to visually create information processing pipelines that combine multiple tools and Iibraries such as VTK, pylab, and matplotlib. We will also show how VisTrails leverages provenance information not only to support result reproducibility, but also to simplify the creation and refinement of pipelines.