Digital Twin Patents Dataset

This dataset contains 396 patent families retrieved from PatBase and functionally classified as part of the study "What digital twin patents protect: A functional interpretation of patent landscapes", submitted to World Patent Information. The dataset supports replication of all empirical results reported in the paper.Coverage: Ten major industrial firms active in digital twin-related patenting - Siemens, IBM, Rockwell Automation, Robert Bosch, Hitachi, ABB, Honeywell, General Electric, Schneider Electric, and Mitsubishi Electric - over the period 2003–2023. Assignee names were harmonised using the OECD HAN database (December 2023 version).Unit of analysis: Patent family (one record per underlying invention, consolidated across jurisdictions using earliest priority date).Classification: Each patent family was assigned to one of three functional categories - Descriptive (D), Predictive (P), or Prescriptive (Pr) - based on a hierarchical rule-based coding procedure combining IPC subclass signals and title-level keyword indicators. Inter-coder reliability was assessed on a stratified validation sample of 90 families (κ = 0.609 pre-adjudication; κ = 0.912 post-adjudication).