DSI Spring Un-seminar
When Data Disagree: Incorporating information from multiple sources into a comprehensive spatial model
presented by Quinn Hart, UC Davis Library Digital Applications Manager
What? This is a different kind of seminar. Come brainstorm about cross-University research problems in this multidisciplinary, flipped series where the audience does most of the talking to apply data science and interdisciplinary knowledge to address complex unsolved research problems. The presenter provides the question(s), and the audience leverages their combined backgrounds and expertise to suggest solutions. Who? All members of the UC Davis research community (faculty, students, postdocs, staff) with a background and/or interest in data science. No experience is necessary to attend - some of the most creative solutions have come from other disciplines. Everybody can gain by participating. Why? These seminars are fun, enlightening, highly multidisciplinary, thought-provoking, and focus on cutting-edge research in progress. By watching problem solving in action you will learn a lot about data science, how to ask the right questions to get the answers you need, and how to extract information from clients and/or your own reserach program.
Satellite imagery and GIS modeling allow us the power to address complex environmental questions over large regions. But, what how do we address issues when these data don’t agree with more local sensors? Can we combine data from multiple sources, with varying resolutions and biases, into a single, comprehensive spatial model? We will discuss a modelling effort being conducted in collaboration with California Irrigation Information Management Information System (CIMIS) program to create new Evapotranspiration maps of California water use zones based on data from the Department of Water Resources (DWR) weather stations and the GOES weather satellite, collected over 13 years. The spatial modeling efforts have led to inconsistencies with some DWR weather stations being assigned to inappropriate zones. How can we best utilize all of the available data and correct for these mismatches? How will we know if the implemented changes are good ones?
- data integration
- geospatial modeling error
To learn more about the DSI's unique un-Seminars, see here.