Comprehensive Freight Demand Data Collection Framework for Large Urban Areas - Sustainable Urban Logistics: Concepts, Methods and Information Systems - page 97

Table 4 Summary of key data gaps
Freight generation data No sources were identified a, b Production

Consumption

Delivery tours Sequence Only Global Positioning System (GPS) data

private vendors.

Location Low level of detail about location.

OD flows Some source but no complete information.

Empty flows No source identified.

Economic characteristics
of participating agents

Shippers Some source can provide some data.

Carriers
Receivers

Spatial distribution/
Location of
participating agents

Shippers Some source identified that can provide some

data

Carriers
Receivers

Network characteristics Travel time and

costs

Only a low level of detail is provided by the
identified data sources

Use restriction
capacity
Traffic volumes

Special choice processes Mode choice Only the Commodity Flow Survey (CFS) provide

some data

b

Delivery time Low level of detail

Mode attributes Some level of detail

Other economic data Production

functions

No sources were identified

Demand
functions

No sources were identified

Good level of detail from Regional Economic
Information System(REIS) and 2002
Benchmark Input-Output Accounts of the
USA
Notes a ITE Trip Generation Manual contain trip rates but no cargo attracted or produced
information
b The Commodity Flow Survey microdata could provide this information. Access to the data is
restricted

Input-Output
technical
coefficients

estimation of freight Origin-Destination (OD) matrices could be reasonably
achieved 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 ). In fact, the truck OD matrix in the New
York City Metropolitan Planning Organization transportation planning model is
the result of such process (List et al. 2002 ). As a result, there is a wide range of
data collection possibilities that differ in the extent of the synthesis that is con-
ducted. At one end of the spectrum, a modeler could undertake freight demand
modeling with a minimal amount of data, which requires the use of a significant
number of assumptions and demand synthesis techniques. In this context, synthesis
techniques could indeed reduce data collection costs, though at the expense of