Workflow-Based Data Science Pipelines for Digital Phenotyping: Enabling Sustainable & FAIR mobile Health Research

The increasing use of mobile and wearable sensing for digital phenotyping has enabled collection longitudinal datasets across diverse health domains. However, data analysis pipelines remain fragmented and ad hoc, limiting sustainability, and adherence to FAIR principles. Existing workflow systems lack tight integration with study protocols and accessibility for clinical researchers. This paper presents the CARP Data Science Pipeline, an extensible, workflow-based analytics framework integrated with the Copenhagen Research Platform. CARP-DSP enables researchers to define, deploy, and execute human-readable data science workflows spanning backend and edge environments. The framework emphasizes sustainable research and FAIR compliance, supporting both technical and non-technical users through a domain-specific language and AI-assisted tooling. The system also incorporates mechanisms for cross-system workflow interoperability and privacy-aware analytics.