Comprehensive Freight Demand Data Collection Framework for Large Urban Areas - Sustainable Urban Logistics: Concepts, Methods and Information Systems - page 94

Network characteristics. All freight models will require a complete specification
of network characteristics, which are required to model how the freight flows are
going to use the transportation system. Traffic counts are also needed for cali-
bration of the model, and in the case of origin-destination synthesis models, to
estimate Origin-Destination (OD) matrices from secondary data.
Special choice processes. In addition, it is likely that a set of models aimed at
studying specific processes are needed. Examples include: freight mode choice,
decision concerning delivery times (which determines the response of the freight
industry to congestion pricing), among others.
Other economic data. These types of data are intended to characterize specific
aspects of freight demand not covered by the previous categories. These include
production and demand functions of commodities in each region, which is infor-
mation needed by spatial price equilibrium models, and technical coefficients of
Input-Output (IO) models.
Table 2 shows a general classification of models that consider their mathe-
matical structure, the flow unit being considered, and the modeling technique or
principle that supports the model. As before, only the main types are considered
for space reasons.
Table 3 presents the summary of the data requirements of the alternative
modeling approaches, taking into consideration the data categories previously
discussed. It is important to note that some modeling techniques are part of a
modeling framework, where an input sometimes is the output of previous steps
that in themselves could be one of the other modeling approaches.
3.1 Data Sources
After outlining the different modeling approaches and categorizing them based on
data requirements, the authors identified the key data sources. In doing this, the
authors used Holguín-Veras et al. ( 2001 ) as the starting point, because it contains a
comprehensive assessment of the modeling alternatives available at the time. The
authors analyzed the data sources available in the United States to assess their
suitability to support a freight demand modeling exercise. The reader is referred to
Holguín-Veras et al. (Holguín-Veras 2010 ) for the detailed analyses that led to
Table 3 .
The analysis revealed that there are major data gaps, shown in Table 4 , that are
not covered by the data available. The fundamental implication is that most of the
data needed for the development of a freight demand model must be collected
from scratch. However, since almost all freight demand modeling exercises entail
one form or another of data or freight demand synthesis i.e., the estimation of data
or freight demand from secondary sources, it is important to discuss the potential
role of synthesis techniques. The reason is that synthesis serves the purpose of
filling gaps in the data collected. Research has shown that, for instance, the