Detect Spatial Clusters of High Acute Malnutrition Rates • wowi

Child acute malnutrition can lead to death if not identified and treated on time. Driven by a combination of diverse factors, it often exhibits spatial variation. The levels of acute malnutrition are commonly measured through surveys that are representative of the area of interest. Results are then used to inform programme responses — to find and treat the affected children. To that end, programme managers require actionable insights on where acute malnutrition is most prevalent; this is essential for prioritising interventions, especially when resources are limited.

wowi - an expression meaning “where” in Elómwè, a local language spoken in central-northern Mozambique - provides a set of convenient utilities for this purpose. It identifies locations across the survey area wherein acute malnutrition is significantly high (hotspots) or low (coldspots), and unlikely to be due to chance alone.

wowi is a wrapper package built on top of the rsatscan package that enables the use of the SaTScan software from within R. While rsatscan provides general-purpose functionality, wowi was thoughtfully designed for acute malnutrition analysis.

To use wowi, you must have SaTScan installed on your machine, along with the mwana R package that is used under the hood to wrangle anthropometric data. The latter gets installed during wowi installation.

Installation

wowi is not yet on CRAN but can be installed through:

remotes::install_github(repo = "tiwowi/wowi", dependencies = TRUE)

What does `wowi` do?

It takes a dataset with GPS coordinates (latitude and longitude), scans for clusters of acute malnutrition — either high or low, depending on the user’s specification — across the survey area, and returns three main outputs: (1) an interactive HTML map displaying the detected clusters (previewed below); (2) a .txt file containing the results; and (3) a table with summary statistics and metadata parsed from the .txt file.

The package is particularly handy when working with datasets that span multiple areas or administrative units, enabling consistent, area-wise detection of hotspots and coldspots. Moreover, it generates additional GIS-based files (e.g., shapefiles) that can be useful for further geospatial manipulation or integration into other mapping workflows.

A glimpse of detected hotspots and coldspots

Red bubble: hotspot; blue bubble: coldspot; Size of the bubble: span of the spot.

A glimpse of the summary table

## # A tibble: 2 × 18
##  survey_area nr_EAs total_children total_cases `%_cases` location_ids   geo   radius span  children n_cases expected_cases observedExpected relative_risk
##  <chr>        <int>          <int>       <int>     <dbl> <chr>          <chr> <chr>  <chr>    <int>   <int>          <dbl>            <dbl>         <dbl>
## 1 District        36            532         104        19 23,24,26,25,3… 13.6… 1.43 … 1.88…      170       4           33.2             0.12         0.085
## 2 District        36            532         104        19 16,20,14,12,1… 13.8… 26.24… 43.5…      258      84           50.4             1.67         4.46
## # ℹ 4 more variables: `%_cases_in_area` <dbl>, log_lik_ratio <dbl>, pvalue <dbl>, ipc_amn <chr>

Shiny App

wowi comes with a lightweight, field-ready and handy built-in web-based application that enables users to upload their data and benefit from the utilities needless to write code or be well versed in R. The app can be launched by running the bellow-given command:

ww_run_app()

Citation

If you use wowi package in your work, please cite using the suggested citation provided by a call to citation() function as follows:

citation("wowi")

To cite wowi in publications use:

  Tomás Zaba (2025). _wowi: Utilities for detecting statistically
  significant spatial clusters of high acute malnutrition rates using
  SaTScan's Bernoulli spatial-scan model_. R package version 1.0.2,
  <https://tiwowi.github.io/wowi/>.

A BibTeX entry for LaTeX users is

  @Manual{,
    title = {wowi: Utilities for detecting statistically significant spatial clusters of high acute malnutrition rates using SaTScan's Bernoulli spatial-scan model},
    author = {{Tomás Zaba}},
    year = {2025},
    note = {R package version 1.0.2},
    url = {https://tiwowi.github.io/wowi/},
  }

Community guidelines

Feedback, bug reports and feature requests are welcome; file issues or seek support here. If you would like to contribute to the package, please see the contributing guidelines.

wowi: Utilities for detecting statistically significant spatial clusters of high acute malnutrition rates using SaTScan’s Bernoulli spatial-scan model