Riding the Spider: A Network-Sampling Framework for Multi-Platform Data Collections

Research on the digital networked public sphere is not only hindered by challenges in data access but also by a lack of common standards for describing and implementing data collection independently of the form of access or technologies employed. These challenges are particularly pronounced in cross-platform research. In this article, we propose a network-sampling framework to conceptualize, implement, and document explorative data collections in a generalizable, readily operationalizable, and interoperable way. Building on the theoretically established components of the networked public sphere, the concept of multilayer networks, explorative network sampling, and legal and technical realities of cross-platform data access, we segment the data collection process into four modules: a Connector, a Parser, a Filter, and a Sampler. This framework enables researchers not only to describe their data collection in a precise and reproducible way but also to follow guidelines on for developing interoperable software implementations of these modules or to propose new modules themselves.