Replication data and code for "National gradients, local heterogeneity: Most settlement-level variation in Census-reported diabetes lies within Australian remoteness categories"

Reproducibility archive accompanying the manuscript "National gradients, local heterogeneity: Most settlement-level variation in Census-reported diabetes lies within Australian remoteness categories" (Alexeev, Gwynne, Henson & Kirwan; submitted to Diabetology). Contains the UCL-level analytic dataset (1,743 Urban Centres and Localities with at least 200 usual residents), a variable codebook, the public ABS Australian Statistical Geography Standard Edition 3 remoteness allocation files, the original Stata and Python analysis/figure code, a standalone Python reproduction check, generated figures, and generated tables/results ledger used for numerical verification. The outcome is Census-reported diagnosed long-term diabetes (2021 Census), not clinical prevalence. Data are derived from Australian Bureau of Statistics 2021 Census TableBuilder output; ABS is acknowledged as the source. Supplementary PHIDU-supplied hospitalisation/mortality material and named-settlement residual diagnostics are not included; see README.md for provenance and exclusions.