Frictionless data is a term pionered by Open Knowledge Foundation with a particular implementacion using CSV and JSON, but the core idea that what matter with data is packaging to enable interaction, instead of size, could be implemented using other combinations like CSV + YAML. My personal preference are human readable formats like YAML and TOML for documentation and metadata, but there is good compatibility among them.
- “Demand carbon dioxide data” says Hans Rosling to open data advocates at OKFestival – page 