Sub-national Analysis with HawaSpatial Pro

Introduction

HawaSpatial Pro (Holistic & Areal Weighted Analysis) is a professional-grade, multi-sectoral platform designed to bridge the gap between complex spatial modeling and actionable policy insights. It provides a robust framework for sub-national monitoring of development indicators, specifically optimized for large-scale household surveys such as the Demographic and Health Surveys (DHS), Multiple Indicator Cluster Surveys (MICS), and Integrated Household Budget Surveys (IHBS).

The package is structured around four pillars of data science: 1. Descriptive Analytics: Establishing baselines and measuring geographic concentration (Gini/Theil). 2. Diagnostic Analytics: Identifying hotspots via 17 ESDA methods and spatiotemporal dynamics. 3. Predictive Modeling: Stabilizing sub-national estimates using Small Area Estimation (SAE) and Multilevel Modeling (GLMM). 4. Prescriptive Analytics: Identifying SDG target gaps and generating formal technical reports.

Getting Started

Data Requirements

To utilize the full power of HawaSpatial, users typically require: * A standard survey dataset (.dta, .sav, or .csv). * A corresponding administrative shapefile (Admin1 or Admin2). * The Spatial Equalizer module within the app will automatically handle naming inconsistencies between these two sources.

Launching the Application

The primary way to interact with the HawaSpatial engine is through its interactive dashboard. To launch the application, use the following command:

library(HawaSpatial)

# Launch the interactive dashboard
run_app()

Core Modules

1. The Spatial Equalizer

The platform solves the common “naming mismatch” problem between survey data and GIS boundaries. It uses a smart-matching logic to join survey statistics to spatial geometries automatically.

2. 4-Stage Hierarchical Modeling

HawaSpatial implements a formal 4-stage multilevel modeling workflow (Null, Individual, Community, and Full models) using the glmmTMB engine. It automatically calculates the Intraclass Correlation (ICC) and Proportional Change in Variance (PCV).

3. Small Area Estimation (SAE)

For regions with small sample sizes, the package provides Fay-Herriot area-level models to produce stabilized Empirical Best Linear Unbiased Predictors (EBLUPs), ensuring that sub-national policy decisions are based on statistically sound estimates.

4. Automated Reporting

One of the key features for researchers is the ability to export formal, publication-ready side-by-side comparison tables and comprehensive technical reports in Word, PDF, or HTML formats.

Citation

If you use this package in your research, please cite it as:

Muse, A. H. (2026). HawaSpatial: Holistic and Areal Weighted Analysis for Global Development (Version 0.1.1). Zenodo. https://doi.org/10.5281/zenodo.18290101