Frequently Asked Questions (FAQ)
What is the purpose of this platform?
The platform provides real-time forecasts of dengue risk at both the provincial and Bangkok district level in Thailand. It combines official health data, environmental factors, and machine learning models to assist public health decision-making.
How are the predictions generated?
Predictions are generated using multiple models, including LSTM neural networks and gradient boosted trees trained on historical case data, weather patterns, and population dynamics. A consensus model then blends the outputs for better accuracy.
Where does the data come from?
Administrative boundaries are sourced from GADM (version 4.1). Case data is collected from national and regional health agencies. Subdistrict-level data in Bangkok is refined through university student contributions and field validation efforts.
What is the role of students?
University students across Thailand help validate data quality, geocode district boundaries, and contribute to crowdsourced epidemiological insights. Their contributions feed directly into improving model resolution.
Can the forecasts be downloaded?
At this stage, public access is limited to visualization. However, requests for raw model output or CSV summaries can be made through the contact form for academic or governmental purposes.
What do the different metrics mean?
Each province or district displays values such as predicted cases, incidence rates, and risk classification. The “Consensus Model” metric aggregates predictions from all models for a balanced risk estimate.
Who maintains the platform?
The platform is developed and maintained by ISGlobal with contributions from researchers at Mahidol University, Chiang Mai University, Thammasat University, and other institutions.