Extracting Human-Level Knowledge from Big Numerical Spatial Data

"Big data" has been thrust into the public consciousness as of late, carrying with it sometimes extravagant promises of innovation and benefits for everyone. Certainly there appears to be no shortage of data in sight, yet following through with the promise and potential of big data poses a number of challenges. One significant roadblock to making use of big data is the time and expertise required to perform analysis and derive products from the data. The more the process of extracting spatial knowledge based on human-level concepts, e.g., lakes, valleys, and relations, e.g., near, after, can be automated through the use of computational data analysis, the more time there is for humans to ask important questions and draw conclusions from the data. Here a methodology for extracting human-level knowledge from spatial data is presented alongside an automated means of encoding the derived products into Linked Data for use by a question answering system. The methodology is extensible since it provides a spatial framework that can be used to aggregate any new additional knowledge that may become available as new techniques for extracting knowledge from spatial data is made available.