Roman Lukyanenko

Citizen science 2.0: data management principles to harness the power of the crowd

Citizen science refers to voluntary participation by the general public in scientific endeavors. Although citizen science has a long tradition, the rise of online communities and user-generated web content has the potential to greatly expand its scope and contributions. Citizens spread across a large area will collect more information than an individual researcher can. Because citizen scientists tend to make observations about areas they know well, data are likely to be very detailed. Although the potential for engaging citizen scientists is extensive, there are challenges as well. In this paper we consider one such challenge - creating an environment in which non-experts in a scientific domain can provide appropriate and accurate data regarding their observations. We describe the problem in the context of a research project that includes the development of a website to collect citizen-generated data on the distribution of plants and animals in a geographic region. We propose an approach that can improve the quantity and quality of data collected in such projects by organizing data using instance-based data structures. Potential implications of this approach are discussed and plans for future research to validate the design are described.

Unintended consequences of class-based ontological commitment

Domain ontologies can promote shared understanding of a domain using standardized constructs to represent domain semantics. Building upon this promise, libraries of domain-specific ontologies have been created. However, we argue that adopting any particular domain ontology can in fact hinder domain understanding and reduce the quality of data made available through a shared ontology. This paper examines unintended consequences of class-based ontological commitment and advocates instead an instance-andproperty ontological foundation. The proposed approach can better inform the practice of information sharing and can, in participative domains, enable users to contribute higher quality information with fewer constraints.