Using Analytics for Community Monitoring and Support in Online Citizen Science Projects

While the body of knowledge about WHAT is learned through participation in online Citizen Science projects has grown over the past years (Aristeidou & Herodotou, 2020), the research field also shifts focus to better understand HOW task-sharing and learning happen in Citizen Science (CS). Technology-enhanced and online citizen science projects allow for analytical tools to be applied directly to digital traces (e.g. Herodotou et al. 2020, August et al., 2020). Even in the absence of explicit (inter-)action logs from project activities, we can rely on accumulated “knowledge artefacts” as data sources, for instance in the form of forum postings or blog entries. Question-answer or request-reply structures in such user-generated knowledge bases can be extracted and transformed into social network graphs that represent the structure of the underlying interactions. Centrality measures such as Eigenvector centrality or “Page Rank” allow for identifying influential users in such contexts (cf. Franceschet, 2011; Tang & Yang, 2010). The distribution of these measures of influence between citizen scientists (volunteers), assigned community moderators and professional scientists reflects the roles of these different groups in a project community. Increasing individual centrality values over time indicate a growing influence of a participant and enable us to map individual learning trajectories.

In the context of the EU project CS Track, we have applied this methodology to the popular Zooniverse project Chimp&See, which is based on the analysis of wildlife camera recordings especially of chimpanzees across Africa. The project has an explicit conservation goal (SDG 15 - Life on land) and has involved more than 5500 volunteers. We analysed the communication between volunteers, scientists and moderators in the public discussion forum using techniques of social network analysis. The findings show that moderators play a crucial role in mediating and coordinating citizen science activities. Using this example, we demonstrate the potential of network analysis methods to help in the design, facilitation and assessment of participation, decision-making and knowledge-building in such online communities. This has the potential to support CS projects in contributing to quality education (SDG 4) and possibly even gender equality (SDG 5) in science education and participation in science.