Modeling and Communicating the Conceptual Intent of Geo‐Analytical Tasks for Human‐GISInteraction

Abstract

AbstractOne of the fundamental issues of geographical information science is to designGISinterfaces and functionalities in a way that is easy to understand, teach, and use. Unfortunately, current geographical information systems (includingArcGIS) remains very difficult to use as spatial analysis tools, because they organize and expose functionalities according toGISdata structures and processing algorithms. As a result,GISinterfaces are conceptually confusing, cognitively complex, and semantically disconnected from the way human reason about spatial analytical activities. In this article, we propose an approach that structuresGISanalytical functions based on the notion of “analytical intent”. We describe an experiment that replacesArcGISdesktop interface with a conversational interface, to enable mixed‐initiative user‐system interactions at the level of analytical intentions. We initially focus on the subset ofGISfunctions that are relevant to “finding what's inside” as described byMitchell, but the general principles apply to other types of spatial analysis. This work demonstrates the feasibility of delegating some spatial thinking tasks to computational agents, and also raises future research questions that are key to building a better theory of spatial thinking withGIS.