Forecasting Disease Outbreaks: Data Science for Epidemiology

Please join Charlottesville Data Science for a double-feature exploring the role of data science in modern epidemic forecasting. Aniruddha Adiga, a Research Assistant Professor at the UVA Biocomplexity Institute, and VP Nagraj, a practicing data scientist and Ph.D. candidate at the UVA School of Data Science, will discuss the evolving methods, data sources, and technical infrastructure shaping how we predict and respond to infectious disease outbreaks.

→ Are you signed up for the Charlottesville Data Science Substack? If not, please consider subscribing at https://news.cvilleds.org. It's free, and it's the primary way we share announcements and event updates with the data science, AI, and machine learning community in Charlottesville and Central Virginia. ←

Event Overview

The COVID-19 pandemic has underscored the need for robust disease surveillance systems and outbreak analytic tools. While infectious disease modeling has been an established domain for decades, the available data streams, computational techniques, and expectations from public health stakeholders have evolved rapidly since 2020. One of the most notable shifts is the operational focus on collaborative modeling "hubs" and ensemble techniques. Currently, there are dozens of organized efforts in the US and abroad to operationally consolidate near-term forecasts and long-term projections of disease targets.

This double header will feature talks from two Charlottesville-based data scientists who are leading operational disease modeling projects and actively contributing to collaborative hub efforts. The speakers will share their experiences navigating the changing landscape of epidemiological data sources and forecasting initiatives. Each talk will combine big picture context with details on technical challenges and methodology.

Talk 1: Real-time Forecasting of Infectious Diseases: Challenges and Innovations

Aniruddha Adiga will discuss key challenges in real-time infectious disease forecasting and ongoing community efforts to develop robust modeling frameworks. He will outline the range of modeling approaches currently used in the forecasting community, highlighting their respective strengths and limitations. The talk will also showcase novel methods developed by the team at the Biocomplexity Institute, University of Virginia, aimed at improving forecast robustness and integrating diverse auxiliary data sources.

Talk 2: Automated Infectious Disease Modeling: Why? How? Should?

VP Nagraj will discuss practical considerations for automating infectious disease model training and inference. Since 2021, he has helped develop and refine cloud-based approaches to automatically prepare forecasts. He will present details of the AWS infrastructure, containerization techniques, and event-driven pipelines that have been implemented for these tasks. While presenting the technical specifics, the talk will address the tension between advantages and pitfalls of attempting to automate operational disease modeling activities.

Getting to the event

The event will take place in the UVA School of Data Science building at 1919 Ivy Road. We'll be in the Capital One Hub on the ground floor. Parking in the adjacent Emmet/Ivy Parking Garage is free after hours.