Open Selection Graph (OSG): a dataset for mapping observable populations and selection processes in science

Open Selection Graph is a dataset for studying how candidates move through observable selection processes in science. It brings together public evidence from peer review, research funding and patent examination. The common unit is an institutional cycle: venues, tracks, calls, panels, journal periods, or examination processes governed by a defined set of rules. Version 2.0.0 contains six licence-separated data components, a code and metadata bundle, the data descriptor package and a checksum manifest. Package-level terms and source-specific licences are recorded in each archive. The release includes the full privacy-minimised 4,518,254-row HUPD census and a separately identified 8,143-case PANORAMA process sample with complete reconciliation. Contact: Anton Waagaard, Dimensional Impact Lab, [e-mail].