Comprehensive Freight Demand Data Collection Framework for Large Urban Areas - Sustainable Urban Logistics: Concepts, Methods and Information Systems - page 108

6 Conclusions
This chapter discusses the key characteristics of the freight system and the data
requirements of the different modeling techniques to design a comprehensive
freight data collection framework for mid-size and large urban areas. In doing so, a
number of findings of great importance are produced. First, the freight system is
comprised of a set of interacting economic agents along the supply chain that
produce, process, transport, distribute, or consume goods. As a result of their
different roles, no single agent can provide a complete picture of the freight
system. This complicates data collection, as assembling a description of the system
requires collecting and putting together the information that each agent is aware of.
Adding to the complexity is the fact that there are different ways to measure freight
and its generation, including: value of the cargo, amount transported, vehicle trips
produced, number of stops and deliveries made and received.
Second, data collection efforts must also take into consideration the data
requirements of the different freight demand modeling techniques. Characteristics
in terms of: the fundamental structure and empirical foundation of the model, the
computational algorithms and data structures that perform the computations, and
the process to analyze the modeling results. In addition, there are requirements for
model development, model calibration and forecasting.
Third, an analysis of the different data sources showed that there are major data
gaps. As a result, the data needed for the development of freight demand models
must be collected practically from scratch. However, some data could be estimated
from secondary sources, using data or synthesis. Data synthesis has been suc-
cessfully used to estimate freight OD matrices using secondary sources such as
traffic counts (Tamin and Willumsen 1988 ; Gedeon et al. 1993 ; Tavasszy et al.
1994 ; Nozick et al. 1996 ; List et al. 2002 ; Rios et al. 2002 ; Al-Battaineh and Kaysi
2005 ; Holguín-Veras and Patil 2007 ; Holguín-Veras and Patil 2008 ).
Fourth, the complexity of the freight system and the modeling data require-
ments determine the collection technique and data collection approach to be used.
Considering that no single data collection approach can provide a good repre-
sentation of all data categories, it seems clear that a comprehensive approach
composed of a combination of methods may be required.
In order to fulfill the objective of the chapter, a modular set of data collection
approaches is proposed. The alternatives range from no data collection and use of
data synthesis, to extensive data collection efforts. Obviously, while no collecting
data minimizes costs; this is at the expense of the modeling effort's quality. In
essence, each alternative has advantages and disadvantages, and as typical of
situations like this, the best approach would be the combination of alternatives that
best fits the needs and constraints of the participating agencies.
Acknowledgments This research was supported by the New York Metropolitan Transportation
Council's project ''Feasibility Study for Freight Data Collection''. The authors also acknowledge
the contributions from Ms. Lisa Destro. The support and contributions are both acknowledged
and appreciated.