Collaborative Research: Frameworks: Automated Quality Assurance and Quality Control for the StraboSpot Geologic Information System and Observational Data

Oregon State University, Corvallis OR

Investigators

Abstract

Digital collection and storage of geological data allow for it to be findable and accessible. For the data to be reusable and reproducible, however, there needs to be some evaluation of the quality of the data and/or trust in the person who collected it. This project focuses on developing an automated system for evaluating the quality of collected data and the completeness of observations. This includes the development of two methods - one based on machine learning and one on logic derived from experts - to evaluate data quality. Furthermore, the project involves testing methods in two areas – field- and lab-based data – to ensure that the methodologies apply to different data types. A mechanism will also be provided that allows scientists to evaluate existing databases and improve data collection while it is occurring. These assessments are essential for expert users to reuse and reproduce observations and for the general public and non-disciplinary experts to recognize high-quality and complete data collections. This project addresses the transformational task of providing an automated QAQC (Quality Assessment/Quality Control) system for observationally-based geological data. The approach integrates Computer Science, Cognitive Science, and Geology expertise to develop algorithms to implement a QAQC system. The basis of these CI resources is the StraboSpot geologic information system. Project activities include the development of two complementary algorithms – one based on machine learning and one on logic derived from experts – to evaluate observational data. Expert testing will be done to improve algorithmic performance. A GUI (Graphical User Interface) will also be developed to allow geology practitioners to evaluate others’ datasets and improve their own during data collection. This approach enables new kinds of science, including: 1) empowering modeling at a regional scale beyond which a single geologist or team could achieve through their use of trusted shared data; and 2) allowing experts in one area to incorporate data outside of their expertise into a model or interpretation, without having to learn how to collect the data and or assess someone else’s data quality. Thus, the proposed work will promote and facilitate using shared data sets within and between disciplines. The iterative, collaborative process through which the QAQC system is designed will serve as a community-building endeavor. This award by the NSF Office of Advanced Cyberinfrastructure is jointly supported by the Division of Research, Innovation, Synergies, and Education (RISE), the Division of Earth Sciences (EAR), and the Tectonics Program within the NSF Directorate for Geosciences. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.