Spatial Data Relations as a Means to Enrich Species Observations from Crowdsourcing

Abstract

The general fascination of nature has always been a major driver for studies on living animal and plant species. A large number of professionals and especially volunteers are organized in related initiatives and projects from the local to the global level, leading to the vast amount of species observations nowadays available on the Web. This article seeks to enhance this knowledge base by the determination, management and analysis of feature entity relations among the observations. Those relationships are considered important for comprehensive biological monitoring and, in general, facilitate the integrated use of existing data sources on the Web. Particular emphasis is put on crowdsourcing, which increasingly receives attention and support by citizen science initiatives. The Linked Data paradigm, representing the core of the Semantic Web, is applied to describe, handle and exploit relations in a standardized and thus interoperable manner. Methodologies to determine and validate relationships are developed and implemented. The implementation combines the analysis of spatio-temporal behavioral patterns of species with a crowdsourcing approach for the validation of determined relations. The vagueness of results is addressed by assessing the probability of a relation.