A cross-platform spatio-temporal visualisation web-application for infectious disease forecasting

This project is part of the Dengue Advanced Readiness Tools (DART) initiative, developed in collaboration with Oxford University Clinical Research Unit (OUCRU). It focuses on building a cross-platform web-based dashboard for visualising dengue forecasting data and supporting back-end system to monitor, analyze, and alert stakeholders about the dengue fever situation in Ho Chi Minh City. Our intention is to provide health authorities and relevant parties with cases incidence, forecasts data, patterns, and real-time alerts to support timely responses. Two main components in this project include: a Dashboard Visualization Tool and a question-answering Large Language Model (LLM) for infectious diseases.



Targeting a wide base of audiences, three roles will be implemented in the system representing the general public, researchers, and decision makers, with a clear separation between the public's generalised case reports and internal detailed forecasting data for the latter two. As for the chatbot LLM, we propose the use of Low-rank Adaptation (LoRA) method, using medical articles as a data source to fine-tune Llama-7B.