A survey of methods for data inclusion in system dynamics models: Methods, tools and applications

Abstract

Numerical data is experiencing a renaissance because 1) traditional data such as census and economic surveys are more readily accessible 2) new sensors are measuring things that have never been measured before, and 3) previously 'unstructured' data - such as raw text, audio, images, and videos - is becoming more amenable to quantification. Because of this explosion and the popular buzz surrounding ‘Big Data’, clients expect to see strong incorporation of data methods into dynamic models, and it is imperative that System Dynamics Modelers are fully versed in the techniques for doing so. The SD literature contains surveys that explain methods for including data in system dynamics modeling, but techniques have continued to develop. This paper attempts to bring these surveys up to date, and serve as a menu of modern techniques.