Online Demonstrations

Defining a Web

Using a Web

SciWalker

The Internet has opened up the opportunity for scientists to access datasets from geographically distributed locations. While this creates opportunities for scientists to more easily build upon the work of others, it also creates risks that the accessed datasets will be used in inconsistent or unsound ways. Datasets used in scientific processes may be far removed from the raw data originally collected in the laboratory or in the field. These synthesized datasets are often produced using common processes, such as sampling, data checking and cleaning, statistical model construction, and statistical inference and evaluation. Scientific consumers of these synthesized datasets, however, need to know precisely what processing has been done so that they can avoid subsequently performing redundant or inconsistent processing on these datasets.

An analytic web is a collection of related graph structures that taken together provide precise definitions of scientific processes used to process raw and derived datasets. The formalisms used to define analytic webs are adaptations of those used in software engineering and provide the basis for automated synthesis and validation of scientific processes and their resulting datasets.

SciWalker is a toolset that supports the creation, execution, and validation of scientific processes represented as an analytic web. Currently SciWalker provides support for creating and editing Data Flow Graph (DFG) descriptions of a process, for binding DFG nodes to specific processing tools and data sets, for executing a DFG, and for creating a Data Derivation Graph (DDG) that completely describes the processes used to derive datasets.

The SciWalker prototype is available for download and use. Please contact us for more information.

Features

• Uses the internet standard WebDAV for data storage
• Supports i ntegration with R for statistical processing
• JavaScript support for the specification of simple activities
• Extensible architecture supports integration with other tools
• Support for the creation and visualization of analytic web graphs

Selected Publications

 2006

Analytic Webs Support the Synthesis of Ecological Data Sets

Aaron M. Ellison, Leon J. Osterweil, Julian L. Hadley, Alexander Wise, Emery Boose, Lori A. Clarke, David R. Foster, Allen Hanson, David Jensen, Paul Kuzeja, Edward Riseman, Howard Schultz, Ecology, Vol. 87, No. 6, pp. 1345-1358, June 2006. (UM-CS-2004-079)

This is an updated version originally dated October 2004

[ Ecology ] [ PDF ]

 2005

Process Technology To Facilitate the Conduct of Science

Leon J. Osterweil, Alexander Wise, Lori Clarke, Aaron M. Ellison, Julian L. Hadley, Emery Boose, David R. Foster, Software Process Workshop (SPW2005), Beijing, China, Springer-Verlag Lecture Notes in Computer Science, Vol. 3840, pp. 403-415, May 25-27, 2005. (UM-CS-2005-017)

[ SpringerLink ] [ PDF ]

© 2006 University of Massachusetts Amherst. Site Policies.
This site is maintained by the Laboratory for Advanced Software Engineering Research.