Vega: A Flexible Data Model for Environmental Time Series Data
Luke Winslow, Barbara Benson, Kenneth Chiu, Paul Hanson, Timothy Kratz
Last modified: 2008-08-21
As large scale sensor networks grow, effective data curation of large data volumes is becoming important. Many sites have filled this need with site-specific database systems and software. Within the Global Lake Ecological Observatory Network (GLEON), a fundamental need for a data model allowing for growth and flexibility in sensing platforms and configurations requiring minimal or no data model changes was identified. The Vega data model is designed to fulfill that need. The Vega data model is a flexible, site agnostic data model optimized for high temporal resolution ecological sensor network data sampled at frequencies as high as a few seconds. Instead of storing data in a spreadsheet-like view with different variables denoted by columns, Vega stores observed values individually and describes them fully with linked, metadata containing tables. While being difficult to intuitively recognize, this more flexible and portable data model is beneficial at the individual institution level because it handles additional sensor deployments and configuration changes with no change in structure and at an inter-institution level because it represents a portable standard against which flexible and site agnostic software can be developed. Deployment and testing of this system has already begun within GLEON and has involved five different institutions.