Internet data-mining of natural history


ITHACA, N.Y. -- Leading-edge computer technologies for organizing, analyzing and disseminating large natural history data sets will be developed at Cornell University, with a $2.6 million grant from the National Science Foundation (NSF) to the Cornell Laboratory of Ornithology, in partnership with the university's Department of Computer Science.

The four-year project is expected to produce new techniques for interactive exploration and analysis of massive, spatio-temporal data collections. Natural history collections represent one of the largest, longest-running -- and virtually untapped -- data sets in existence.

"These data hold the answers to many biological questions that cannot be answered without more sophisticated tools to organize and analyze the information," says Steve Kelling, the Lab of Ornithology's information technologies (IT) director and a principal investigator (PI) on the grant. "The funding provided by NSF will allow us to build those tools by bringing together the expertise of computational specialists, population biologists and statisticians, with the goal of making the data sets accessible to researchers, conservation biologists, educators and citizens."

Kelling's co-PIs at the Lab of Ornithology are Director John Fitzpatrick and Assistant Bird Population Studies Program Director Wesley Hochachka. Rich Caruana, an assistant professor of computer science whose research is in machine learning, and Mirek Riedewald, a computer science research associate whose research is in data mining, also are co-PIs on the grant.

The grant is one of several the Cornell Lab of Ornithology has received from NSF for its efforts to develop Internet-based data collection and dissemination. In recent years, the lab's researchers were granted $2.25 million for Citizen Science Online, an initiative to develop online tools to engage the general public, including schoolchildren, in the collection of bird observation data. One project resulting from that effort, eBird, an online checklist program that allows bird enthusiasts to report bird sightings anytime over the Internet from anywhere in North America, received additional NSF funding last year. eBird now reaches into Mexico.

The lab also received funding to bring to the Web Birds of North America , an 18-volume compilation of the life histories of all North American bird species. Thanks to the NSF funding for BNA Online, researchers, students, birders, teachers and others have access to perpetually updated species accounts, based on the latest scientific findings. The accounts include vocalizations and video clips of bird behaviors provided by the lab's Macaulay Library.

Says computer scientist Caruana: "This grant will allow us not only to develop and test leading-edge data mining, machine learning and modeling techniques, but to integrate these methodologies with the ornithology lab's existing projects. The project has the potential to greatly increase the power and scope of these data tools and the ability of researchers everywhere to better understand what the natural history data are telling us."

Caruana says the plan to develop new data mining techniques that will allow modeling of long-term population trends, geospatial abundance patterns, and seasonal or irruptive species movements will result in a revolutionary infrastructure for the environmental community and beyond.

Source: Eurekalert & others

Last reviewed: By John M. Grohol, Psy.D. on 21 Feb 2009
    Published on All rights reserved.