DATA MODELING FOR BUSINESS INTELLIGENCE IN TRANSPORTATION LOGISTICS

Resumo

This chapter aims to lay the conceptual and technical groundwork needed to build a Business Intelligence (hereafter, BI) environment for transportation logistics. The goal is not to offer a superficial review of generic information technology concepts artificially transplanted into the transportation industry. Rather, drawing on the established literature on data architecture and dimensional modeling, it builds a framework that speaks directly to the operational realities of freight and passenger transportation—namely, the dispersion of data sources, the heterogeneity of data formats and capture frequencies, and the structural fragility of integration between legacy systems and enterprise resource planning (ERP) platforms. The chapter covers five complementary areas. First, it discusses why transportation logistics is a particularly complex data domain, in light of the information systems and supply chain literature (CHRISTOPHER, 2016; BALLOU, 2006). Second, it presents a layered data architecture—from the operational source to the analytical dashboard—grounded in the seminal contributions of Inmon (2005) and Kimball and Ross (2013) on data warehousing and dimensional modeling. Third, it details the Extract, Transform, and Load (ETL) processes, with particular attention to the data quality issues discussed by Wang and Strong (1996) and Redman (2001). Fourth, it examines integration with enterprise management systems and the viable strategies for dealing with legacy systems that lack a structured application programming interface. Finally, it consolidates best practices in data modeling and data governance, aligned with the DAMA International (2017) body of knowledge, and presents a real, anonymized case study that illustrates the practical application of the concepts discussed. The decision to address these topics in an integrated way—rather than as isolated blocks of technical content—reflects an underlying conviction that guides this entire chapter, one shaped by the author’s hands-on experience building fleet control departments and implementing quality management systems (ISO 9001 and SiAC/PBQP-H): BI projects fail far more often because of flawed architectural decisions at the outset than because of limitations in the visualization tools used at the end of the process. A visually sophisticated dashboard built on a poorly modeled database is a façade—and that façade is exactly what this chapter sets out to prevent.