Introduction and course overview

Introduction

This is the Forecasting & evaluation course, designed for SISMID in 2026. The companion Nowcasting & \(R_t\) estimation course covers additional content on delay distributions, the renewal equation, and nowcasting.

Slides for this session

Aim of the course

In this course we will address how we can use data typically collected in an outbreak, or in routine surveillance, to answer questions like

  • What does the recent trend mean for the near future (forecasting)?
  • How good are our predictions, and how can we tell (forecast evaluation)?
  • How can we combine predictions from multiple models (ensembles), and contribute them to a shared forecasting effort (collaborative modelling / hubs)?

To answer these questions, we need to understand the kinds of data that we typically have available for outbreak analysis and infectious disease surveillance and the principles of good probabilistic prediction. Ideally, we’d also layer in some understanding of the epidemiological processes that give rise to outbreaks.

There are particular challenges when trying to do these analyses in real time (i.e. whilst transmission and data collection is ongoing) rather than retrospectively, which we will address in turn.

In this course, we focus on predicting the future (forecasting), evaluating how good those predictions are, combining models into ensembles, and contributing forecasts to collaborative modelling hubs.

Why this course?

  • Forecasts are becoming increasingly sought-after during outbreaks
  • It’s easy to make bad forecasts and hard to make good ones: this class will teach you how to distinguish between them and a bit about best practices for building epidemic forecasts
  • There’s currently (at the time of devising this course) no comprehensive training resource that links these common questions and challenges

Approach

Throughout the course we will

  1. work with real epidemiological surveillance data in R
  2. fit time-series forecasting models, make predictions, and evaluate them using proper scoring rules
  3. combine models into ensembles and contribute forecasts to collaborative modelling hubs

Each session in the course:

  • builds on the previous one so that participants will have an overview of the real-time analysis workflow by the end of the course;
  • starts with a short introductory talk;
  • mainly consists of interactive content that participants will work through;
  • has optional/additional material that can be skipped or completed after the course ends;

For those attending the in-person version the course also:

  • has multiple instructors ready to answer questions about this content; if several people have a similar question we may pause the session and discuss it with the group;
  • follows a stop-and-review approach where we pause after each section of self-guided material to discuss and review together and address any questions;
  • ends with a wrap-up and discussion where we review the sessions material.

Timeline for the course

This forecasting & evaluation course runs over the second half of the SISMID week (Wednesday midday to Friday), but of course if you are studying this on your own using the web site you can go through the material at your own pace and in your own time. Broadly, the intended order of content is:

  • forecasting concepts and models
  • forecast evaluation and ensembles
  • collaborative forecasting with hubs and course wrap-up

To get started: