HydroCloud: An Online Integrative Tool for Hydrologic Data
Mike McGuire (Computer and Information Science) & Martin Roberge (Geography and Environmental Planning)
The need to analyze and predict the weather is important enough that a bewildering array of hydrologic data products have arisen to meet this demand. Surprisingly however, many of these products do not get used in conjunction with one another, despite their great synergistic potential. In part, this is because the data sets are large, and they are stored in different locations, in different formats, by different agencies. In this project, we propose to develop a prototype, distributed, cloud-based data warehouse architecture for integrating multi-scalar, spatial and temporal hydrologic data. The “HydroCloud” system will use a cloud computing framework that distributes processing and data storage over a connected set of virtual servers. This prototype data warehouse system will include the ability to provide online analytical processing (OLAP) to allow hydrologists to analyze data across a number of spatial, temporal, and attribute dimensions. Project objectives include:
Develop data-driven hydrologic research questions based on real-world hydrologic applications. This will include the identification of hydrologic applications that will specifically benefit from our distributed spatio-temporal OLAP approach, as well as the targeting of additional datasets to include in the prototype system.
Analyze requirements for a spatio-temporal data warehousing environment for hydrologic research. The above research questions will be converted to OLAP queries. These queries will then be analyzed to develop formal requirements for an integrated data model.
Design and build a prototype, distributed, cloud-based system architecture for the storage and retrieval of multi-scale spatial and temporal hydrologic data. The prototype system will be based on a problem domain that is tractable in terms of spatial and temporal resolution as well as the number and type of the datasets to be integrated.
Test and evaluate the prototype system. The system will be tested in a real-world scenario based on the data-driven exploratory analysis of hydrologic data.
The resulting system will serve as a proof of concept for a much larger proposal to be submitted to the National Science Foundation in July of 2013.
Impact on Students
Michael P. McGuire, V. Janeja, A. Gangopadhyay. (2012) “Exploring Multivariate Spatio-Temporal Change in Climate Data Using Image Analysis Techniques.” COM.GEO 2012 International Conference on Computing for Geospatial Research, July 1 - 3, Washington DC, USA, 2012
Michael P. McGuire, V. Janeja, and A. Gangopadhyay. (2011) “Characterizing Large Sensor Datasets with Multi-Granular Spatio-Temporal Intervals” ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, November 1 - 4, 2011, Chicago, Illinois.
Michael P. McGuire, Vandana P. Janeja and Aryya Gangopadhyay (2010). Spatiotemporal Neighborhood Discovery for Sensor Data, Lecture Notes in Computer Science, 5840, pp. 203-225. DOI: 10.1007/978-3-642-12519-5_12
Michael P. McGuire, A. Komlodi, A. Gangopadhyay, and C. Swan. (2008). A user-centered design for a spatiotemporal data warehouse for data exploration in ecological research. Ecological Informatics 3(5). 273-285. DOI:10.1016/j.ecoinf.2008.08.002