Analytical methods for Sub-National Tailoring (SNT) of malaria control
strategies. sntmethods turns DHS/MIS survey microdata and routine health
facility data into survey-weighted indicators, model-based geostatistical
surfaces, and incidence estimates for sub-national program planning. It is the
analytical companion to
sntutils (data I/O, cleaning,
dictionaries, spatial validation).
Three workflows, each with a dedicated guide on the package website:
- DHS survey analysis — survey-weighted, design-correct estimates for 100+ indicators across 16 domains, with confidence intervals and admin stratification. Every family ships a machine-readable data dictionary, and you can inspect the available DHS variables before building any indicator.
- Spatial modeling (MBG)
— model-based geostatistics that turns cluster-level survey data into
continuous raster surfaces and population-weighted admin estimates, via
fit_mbg_indicator()(single indicator) orrun_mbg_pipeline()(full pipeline across 14 indicator families). - Routine data: incidence & TPR — malaria incidence from facility data using the N0-N5 cascade, test positivity with structured fallbacks, and STL + Mann-Kendall trend analysis.
All estimators return tidy, long-format tables keyed by admin level (adm0,
adm1, adm2) with point, ci_l, ci_u, numerator, denominator, and
indicator metadata — so survey and modeled estimates stack cleanly.
# install.packages("pak")
pak::pkg_install("ahadi-analytics/sntmethods")Spatial features need GDAL/GEOS/PROJ system libraries, and the MBG workflow
additionally needs INLA and the mbg engine. See
Get started → Install
for the full instructions.
library(sntmethods)
# Discover surveys and inspect variables BEFORE building indicators
ge <- dhs_read(path = "path/to/parquet", file_type = "GE",
survey_type = "DHS", country_code = "TG")
kr <- dhs_read(path = "path/to/parquet", file_type = "KR",
country_code = "TG", survey_year = 2017)
make_dhs_raw_dictionary(kr) # full variable list for the recode
# Compute a survey-weighted indicator
fever <- calc_fever_dhs(dhs_kr = kr, gps_data = ge,
shapefile = shp_admin, admin_level = c("adm0", "adm1"))See the Get started guide
for the full tour, and
inst/scripts/
for complete, runnable examples.
| Guide | Topic |
|---|---|
| Get started | Overview, installation, a short end-to-end tour |
| DHS survey analysis | Survey discovery, variable inspection, 16 indicator domains |
| Spatial modeling (MBG) | fit_mbg_indicator() and run_mbg_pipeline() |
| Routine data: incidence & TPR | N0-N5 cascade, TPR fallbacks |
| Trend analysis | STL + Mann-Kendall + Sen's slope |
| Methodology & conventions | Indicator specs, naming, dictionaries |
| Reference | Every exported function |
Indicator methodology specs (variable mappings, inclusion criteria, WHO/WMR
references) live as machine-readable YAML in
inst/methods/.
- sntutils — companion utilities (data I/O, dictionaries, cleaning, spatial validation)
- DHS.rates — standard DHS mortality rate calculations
- mbg — model-based geostatistics engine (IHME)
MIT