loading page

A Data Library to Archive, Analyze, Visualize and Serve Online Datasets from Multiple Domains in an Interoperable Framework
  • Rémi Cousin,
  • Sruti Devendran,
  • John del Corral
Rémi Cousin
Columbia University of New York

Corresponding Author:remic@iri.columbia.edu

Author Profile
Sruti Devendran
Columbia University of New York
Author Profile
John del Corral
Columbia University
Author Profile

Abstract

The PRISM Data Library (DL) is designed to optimize the display, analysis, and retrieval of multiple domains datasets. Originally created for climate data, we aggregated data from agriculture and hydrology domains, as well as non-traditional domains for the DL such as ecology, finance, power outage and space weather data. These datasets range from simple geospatial point observations, to spatially gridded data products, to high-resolution satellite measurements, to GIS representation of administrative or domain-specific geographic entities. These datasets are represented in a consistent multi-dimensional (most often spatial and temporal) framework. As a result, dimension-wise comparisons are easily enabled through selection or transformation. Gridded data can be averaged over discrete geometrical entities (e.g. Counties, Bird Conservation Regions). The DL can be used in a browser, by connecting to servers at San Diego Supercomputing Center (SDSC) over the internet. Data selection, processing, and analysis are performed by the SDSC DL servers, and the resulting images or data files are sent back to the client’s desktop. This model optimizes the use of internet bandwidth.