Scientific models can help us understand the important features of complex systems, but they need good data.

Predictive modeling uses known results to create, process, and validate a model that can be used to forecast future outcomes.