Abstract
Systems designers are creating configurable overlay systems to allow geospatial decision makers to relate and integrate multiple data sources. Modelers are producing predictive models to help decision makers by relating information directly to decisions, e.g. as “risk surface” overlays. However, these models are inherently bounded in scope, and cannot reflect contextual changes outside of their design. Using a proxy task, we experimentally tested the performance benefits of model products against other common presentation formats, with and without explicit external context rendering the model inaccurate. We examined how participants remediated the impact of context in a configurable system. The results indicated the strengths of access to model products, the weaknesses when provided without the original underlying data, and preferences towards using raw data to adapt to context over other formats. In future work, we will relate the process metrics of interactive behavior when performing the task, to the outcomes reported here.
Get full access to this article
View all access options for this article.