MalariaScan is a web-based platform for real-time detection, analysis, and visualization of spatiotemporal malaria clusters in cross-border region between Brazil and French Guiana.

Built with R Shiny, SaTScan™, and Docker, it enables public health teams to identify potential outbreaks and support data-driven decision-making.

Malaria remains a major public health challenge in border regions such as Brazil–French Guiana, where:

• Surveillance systems are fragmented
• Population mobility is high
• Access to analytical tools is limited

MalariaScan was designed to bridge the gap between advanced spatial analysis and real-world public health needs, making complex methods accessible to non-technical users.

• Detects abnormal increases in malaria cases
• Identifies spatiotemporal clusters using scan statistics
• Supports early outbreak detection
• Provides an interactive and intuitive interface
• Enables data-driven surveillance workflows

• 📍 Interactive maps with Leaflet
• 📈 Time series visualization with Highcharter
• 🔍 Cluster detection via SaTScan™ software
• 🧮 Statistical validation of detected disease clusters
• 🤖 Integrated chatbot with FAQs to guide users in:
  • navigating the application
  • understanding analytical outputs
  • interpreting cluster detection results
• 🧩 Modular R-based architecture
• 🐳 Fully containerized environment via Docker

malariascan/
│
├── app.R                         # Main Shiny application (UI + server)
├── Dockerfile                    # Container setup
├── satscan.10.2.5.linux.tar.gz   # SaTScan™ package used in the Docker build
├── README.md
├── .gitignore
│
├── R/                            # Supporting functions
│   ├── sscan_data.R
│   └── update_prm.R
│
├── data/                         # Input data (not versioned)

• R
• Shiny

• SaTScan™ software

• Docker
• Linux environment

A public demo of MalariaScan can be accessed online throught the following link: https://www.malariascanapp-demo.online/

git clone https://github.com/mairamorenoc/MalariaScan.git
cd malariascan

Create a data/ folder and include:

• Shapefiles
• Malarias cases datasets
• SaTScan .prm configuration file

⚠️ Data is not included in this repository due to size constraints.

docker build -t malariascan .

docker run -p 3838:3838 malariascan

Access the app at: http://localhost:3838

De-identified Malaria Cases Data → R Processing → SaTScan™ → Shiny App → Interactive Dashboard

• Main Components
  • Data layer: Surveillance systems (Brazil's SIVEP-Malária)
  • Data Processing: R (data cleaning & preparation)
  • Statistical Analysis: SaTScan™ (scan statistics)
  • Data Visualization: Shiny, Leaflet, Highcharter
  • Deployment: Docker

The MalariaScan platform integrates prospective spatiotemporal scan statistics to detect and monitor clusters of malaria cases in near real-time.

Cluster detection is performed using the space-time scan statistic implemented in SaTScan™, applying a cylindrical scanning window:

• Base: geographic area
• Height: time interval

This approach enables the identification of spatiotemporal clusters, capturing both the location and timing of potential increases in malaria cases.

The analysis uses the Space-Time Permutation Model, which is particularly suitable for surveillance contexts where:

• Only case data are available
• No reliable population-at-risk data are required
• Data may come from multiple surveillance systems

Instead of relying on expected counts derived from population data, the model:

• Adjusts for purely spatial variation
• Adjusts for purely temporal variation
• Detects unexpected space-time interactions

MalariaScan operates in prospective analysis mode, meaning:

• The system continuously evaluates incoming data
• Only active or emerging clusters are reported
• Clusters are detected as they occur, not retrospectively

This makes the platform suitable for real-time epidemiological surveillance and early warning systems.

The system identifies and reports:

• Geographic location of clusters
• Time interval of occurrence
• Observed vs expected distribution patterns
• Relative concentration of cases
• Statistical significance (p-value)

• Includes full R + Shiny environment
• Designed for reproducibility
• SaTScan must be available inside the container

• Production deployment
• Performance optimization for large datasets
• Alert system for detected clusters
• Expansion to other diseases

• Fundação Oswaldo Cruz (Fiocruz)
• Observatório de Clima e Saúde (Icict/Fiocruz)
• French National Research Institute for Sustainable Development (IRD)
• Conselho Nacional de Desenvolvimento Científico e Tecnológico - Brazil (CNPq)

Key works supporting this project include:

• Kulldorff, Martin, et al. "A space–time permutation scan statistic for disease outbreak detection." PLoS medicine 2.3 (2005): e59.
• Saldanha, Raphael, et al. "Contributing to elimination of cross-border malaria through a standardized solution for case surveillance, data sharing, and data interpretation: development of a cross-border monitoring system." JMIR public health and surveillance 6.3 (2020): e15409.

Maira Alejandra Moreno

[email protected]