Best Of Both Worlds: System Thinking Approach for Transportation Data-Driven Decision-Making

A great interest of Transport Management Centers (TMCs) operators is to use data to make their decisions accordingly. Despite increase in accessible data, current methods has various gaps from imposing strong constraints (e.g., parametric function form) in traditional statistical methods to relying on statistical associations in Machine Learning (ML) tools. Defining causal knowledge from the transportation domain for ML models can potentially overcome those gaps yet it is done implicitly without a formal framework. This interdisciplinary research proposes a Hybrid Dynamical Systems Thinking Approach (HDSTA), using systems thinking for causality interface implementation for data-driven decisions in transportation. HDSTA provide guidelines on how different parties can work together to define a knowledge graph for the transportation system model. The graphical and text description outputs will serve experts in choosing and defining variables' cause-effect relationship; data scientists in defining a causal function; and TMCs in making data-driven decisions for the public benefit.