Drowning in Data: Digital Library Architecture to Support Scientific of Embedded Sensor Networks

C L. Borgman, University of California, Los Angeles
J C. Wallis
Matthew S. Mayernik, University of California, Los Angeles
Alberto Pepe, UCLA

Article comments

ACM/IEEE Joint Conference on Digital Libraries 2007 July 17–23 Vancouver, British Columbia, Canada

Abstract

New technologies for scientific research are producing a deluge of data that is overwhelming traditional tools for data capture, analysis, storage, and access. We report on a study of scientific practices associated with dynamic deployments of embedded sensor networks to identify requirements for data digital libraries. As part of continuing research on scientific data management, we interviewed 22 participants in 5 environmental science projects to identify data types and uses, stages in their data life cycle, and requirements for digital library architecture. We found that scientists need continuous access to their data from the time that field experiments are designed through final analysis and publication, thus reflecting a broader notion of “digital library.” Six categories of requirements are discussed: the ability to obtain and maintain data in the field, verify data in the field, document data context for subsequent interpretation, integrate data from multiple sources, analyze data, and preserve data. Three digital library efforts currently underway within the Center for Embedded Networked Sensing are addressing these requirements, with the goal of a tightly coupled interoperable framework that, in turn, will be a component of cyberinfrastructure for science.