GeoDeepDive

Introduction

GeoDeepDive is a collaboration between Shanan Peters' GeoScience Research group, Miron Livny's Condor Research group, the Center for High Throughput Computing, and the Hazy research group. The goal is to help geo-scientists extract data that is buried in the text, tables, and figures of journal articles and web sites, sometimes called dark data. To bring dark data to light, we use the Hazy research group's machine reading system.

The GeoDeepDive system extends a database created by Professor Peters' group, called Macrostrat. Macrostrat contains information about rock formations and individual geographical units. The units are spatially distributed and temporally distributed over geological time intervals. Macrostrat contains over 36K rock formation units. There is also geologically relevant information in journal articles. GeoDeepDive extracts information about these rock formations from journal articles. As of today, GeoDeepDive has processed over 36K research papers and 134K web pages.

Video Demo

Watch on YouTube

Download (109MB)
Our YouTube Channel


Acknowledgements


Special thanks to NSF for sponsoring this work under EAR-1242902.

GeoDeepDive is also generously supported by gifts from Google. We are also grateful to the generous support from the Condor team and the Center for High Throughput Computing at UW-Madison.