Declarative Approaches to Business Process Deviance Mining: A Systematic Literature Review and Taxonomy — Replication Package

Replication package for the systematic literature review "Declarative Approaches to Business Process Deviance Mining: A Systematic Literature Review and Taxonomy" (PUC-Rio). The package makes the whole review auditable and reproducible in two modes: a deterministic, offline verification mode that recomputes every number reported in the paper from frozen JSON/CSV/XLSX artifacts, and a live re-execution mode that re-runs the retrieval and enrichment pipeline against the source APIs. The review used a hybrid search strategy (database searching over a reproducible bibliographic index plus iterative backward and forward snowballing), executed in June 2026: 1,849 unique records retrieved, 17 primary studies (2007-2026) selected under explicit inclusion, exclusion, and quality criteria. The synthesis contributes D3M, a taxonomy organised around five dimensions, a capability matrix, an inventory of event logs and tools, and a roadmap of seven open challenges. Contents: retrieved and enriched corpora; the first relevance filter (protocol, lexicon and weights, distant-reading inputs, per-record evidence); snowballing state and screening worksheets; the final selection and documented exclusions; the taxonomy dual-coding exercise and the screening/eligibility rater-reliability exercise; the executed retrieval/enrichment pipeline and deterministic verification scripts; the PRISMA 2020 checklist; a taxonomy multi-label robustness analysis; and a replication guide (README-replication.md) reproducing each reported figure. This dataset is the supplementary material of the companion article of the same title (to be linked as a related identifier once the paper receives a DOI).