# PNSCMI Nigeria – data and code for "The 'right facility, wrong place' problem" (Paper 1)

| `pnscmi_lga_results.csv` | One row per LGA (n = 774), keyed by OCHA P-code (`adm2_pcode`). Empirical Bayes–smoothed, survey-weighted NDHS 2023–24 indicators (%): `sba`, `fac_delivery`, `anc4`, `pnc`, `cpr`, `unmet_need`; `n_clusters` (NDHS clusters in LGA); `phni` (Population Health-Need Index; higher = more need); `fac_density` (facilities per 10,000); `hl_density` (secondary + tertiary facilities per 100,000); `sfca_30min` (2SFCA score, 30-min catchment); `hsci` (Health-Service Capacity Index); `pnscmi_score` (z(PHNI) − z(HSCI)); `quadrant`; `lisa` (LISA cluster, p < 0.05); `zone`; `rural` (below national median density); `population` (GRID3 v3.0). Need indicators and indices are missing for the 121 LGAs without an NDHS 2023–24 cluster. || `sfca_all_thresholds.csv` | 2SFCA scores at 15, 30, 45 and 60-minute catchments, with LGA population and demand-point snapping distance (km). || `dhs2018_modelled_lga.csv` | LGA means of the 18 DHS Program 2018 modelled surfaces (proportions, 0–1), used for external and convergent validation. |