Changelog • sntmethods

sntmethods 0.8.0

First public release.

calc_*_dhs() functions produce survey-weighted, design-correct estimates across 16 indicator domains (ITN, IRS, PfPR, ACT, antimalarials, fever care, ANC, IPTp, EPI, SMC, anemia, U5MR, case management, CSB, wealth). All return long-format list(adm0, adm1, ...) tables with confidence intervals, design effects and admin-level stratification.
Every indicator family ships a machine-readable data dictionary (act_dictionary(), itn_dictionary(), …) plus the unified dhs_dictionary().
dhs_read() reads single-survey parquet files directly to preserve value labels and survey-specific variables.
ACT handled as a drug class across multiple variables (combination therapy only; monotherapies excluded).
ITN use_if_access (use among those with access) calculated by default with standard age groups (u5, 5_14, ov15).

run_mbg_pipeline() orchestrates the full model-based geostatistics workflow: survey discovery, cluster-data preparation, INLA/MBG fitting, raster prediction, and population-weighted aggregation to admin-2/3 with roll-up to admin-1 and admin-0.
calc_*_mbg() / prep_*_mbg() prepare cluster-level data; fit_mbg_indicator() fits a single indicator; aggregate_raster_to_admin() performs zonal statistics.
Mapping helpers: generate_all_maps(), generate_indicator_map(), plot_mbg_clusters().
U5MR correctly expressed per 1,000 live births; intervention-coverage indicators as percentages.

calc_tpr() with structured fallback logic and validate_tpr_proxies().
calc_incidence() implementing the N0–N5 incidence cascade, with create_incidence() and check_incidence().
calculate_reporting_metrics(), classify_facility_activity(), get_active_facilities() and detect_outliers() for reporting-rate and data-quality analytics.
run_grouped_stl_trend() for STL decomposition with Mann-Kendall trend tests.

New hex sticker logo at man/figures/logo.png.
pkgdown site and deploy workflow.
Hard dependencies fail fast via rlang::check_installed() for optional (Suggests) packages.