Mapping the (real) world: How causal maps furnish a perfect interface between the world and our models of it.

A crucial challenge in using quantitative risk models to inform decisions is to find an interface that shows the relationship between the model and the real world it represents. A model should provide transparency and insight. Too much mathematics will occlude that clarity for all but the most mathematically articulate, but too little risks turning the model into a black box and abandoning the insight the model was built to provide in the first place.

A causal map or influence diagram illustrates the elements of a model and the fundamental relationships between them. Causality provides a natural language to discuss how interventions – decisions – propagate down causal chains to effect outcomes. A causal map represents mathematical relationships without mathematics, allowing stakeholders to take ownership of the scope of a model, the data that conditions it as well as the metrics that monitor and validate it, without getting tied down in notation.

This presentation will introduce the basic concepts of causal mapping and illustrate with simple examples how they can be used to build risk models.