A Collaborative Platform Supporting Distributed Teams in Visualization and Analysis of Infectious Disease Data

• Main Author: Vogtle F.
• Other Authors: Mahidol University
• Format: Conference or Workshop Item
• Published: 2023
• Subjects: Computer Science
• Online Access: https://repository.li.mahidol.ac.th/handle/123456789/84362

Control of infectious diseases requires insight into transmission dynamics and their relation to relevant spatiotemporal factors. Due to the geographically distributed nature of disease outbreaks, as well as the multidisciplinary teams needed to analyze disease data, the experts needed for analysis and modeling may not all be located in the same place at the same time. There is thus need for an analysis and visualization tool to support distributed teams in upstream and downstream disease modeling tasks. In this paper we present a collaborative platform for visualization and analysis of spatiotemporal data concerning disease incidence and related factors. The platform supports integration of data in a variety of formats and resolutions and creation of derived attributes on the fly. Data can be visualized in terms of 3D choropleth maps, as well as scatter plots which include statistical correlations. Multiple visualizations can be simultaneously displayed and manipulated by all session users. We demonstrate the use of the system with the analysis and modeling of data on dengue incidence and related factors in Thailand. The data includes counts of potential mosquito vector breeding sites extracted from street view images using convolutional neural nets. We show how the visualization supports exploratory data analysis that drives machine learning model development and then show how it helps to understand the model output, which provides insight into how and where the models may be best used.