Modelling practices

An excellent first introduction to how a simple epidemiological model translates assumptions about infection and recovery into a mathematical representation that can be used to reason about epidemic dynamics.

A natural next step for seeing how and why a simple model is extended when additional biological mechanisms or population structure become relevant to the question being addressed.

A broader perspective on how to think critically about a model’s purpose, structure, assumptions, data and interpretation after encountering a first concrete modelling example.

A quick way to resolve one of the most common conceptual confusions in transmission modelling before moving to more technical treatments.

Read it after the short explainer above to understand why alternative transmission functions are not interchangeable and how their assumptions can change model behaviour.

A foundational reference to keep nearby for the mathematical and epidemiological concepts that recur across infectious disease modelling, from compartmental dynamics to stochastic, spatial and network models.

A classic reference for distinguishing generation time, serial interval and related epidemiological quantities that are too easily used interchangeably.

An essential practical read before estimating or interpreting R_t, particularly for understanding how observation delays, interval assumptions and methodological choices can bias estimates.

Useful when planning an inference workflow because it makes the trade-offs between model, simulation method, inference method and software implementation explicit.

A useful bridge between mechanistic epidemic models and likelihood-based inference, particularly when working with stochastic and partially observed systems.

A clear entry point into Bayesian inference for stochastic epidemic models before tackling more specialised or computationally intensive approaches.

Worth reading when the simulator is stochastic and standard likelihood-based fitting becomes difficult, because it clarifies the inferential challenges created by stochastic simulation models.

An accessible introduction to sensitivity analysis, from simple local and one-at-a-time approaches to widely used global methods such as Morris, Sobol and FAST/eFAST, with practical guidance on their respective uses and limitations.

A useful next read for understanding why sensitivity analysis should explore the input space globally, and why apparently simple approaches can miss interactions, nonlinearities and important sources of uncertainty.

A concise next step for understanding why uncertainty in model parameters and predictions must be considered explicitly when models are used to compare actions or support epidemic management decisions.

Especially valuable for animal-health modellers because it shows how model structure, scale, data and complexity should be shaped by the decision the model is intended to support.

A concise demonstration that apparently minor choices about model structure can materially change predicted intervention effects and therefore management conclusions.

A useful reality check before choosing an individual-based model, showing both what individual-level detail can add and the additional complexity it brings.

A short and durable guide for setting up a computational project so that scripts, data, results and experiments remain understandable months or years later.

One of the best compact guides for improving everyday computational practice through pragmatic habits for organising data, writing code, collaborating and making research reproducible.

A compact checklist for turning research code that merely runs into code that is easier to test, maintain, review, share and reuse.

A useful hands-on companion to broader reproducibility principles, with concrete practices for project structure, version control, code quality, dependencies and archiving.

A gentle introduction for readers who first need to understand why version control matters before learning Git commands and workflows.

A concrete checklist for making simulation studies easier to understand, assess and reproduce, particularly for agent-based, discrete-event and system-dynamics models.