Environmental Engineering Reference
In-Depth Information
T
abl
e
8
S
a
mplin
g
alter
n
ative
s
Data
Alternatives
Case 0:
NDC. Use
generation rates
Case 1:
5 surveys per county for each
freight; 10 total for non-freight industries
Case 2:
10 surveys per county for each
freight; 25 total for non-freight industries
Case 3
: 25 surveys for each freight; 5 for
each non-freight industries per county
A: Small investment
A: Support industry pooled models only
A: Support industry pooled models with
county parameters
A: Support models by inidustry
classification and county, pooled if desired
L
: No connection to
local conditions
L
: No ability to consider county specific
models
L
: No ability to have county specific
models by industry
L
: None
Case 0: NDC. Origin-
Destination
Synthesis
Case 1: Purchase sample from
GPS data aggregators
Case 2: 1% of commercial
vehicle registrations
Case 3: 2% of commercial
vehicle registrations
Case 4: 3% of commercial
vehicle registrations
A
: Small investment
A
: Lowest cost
A
: Data appropriate for
modeling. Low cost
A
: Data with less gaps.
Appropriate for modeling
A
: Modeling needs likely met.
Meets OD survey standards.
L: Weak/No trip
determinant data
L: Potentially large bias in data.
No trip determinant data
Case 1:
1 day of data collection
L: Some data gaps may be
evident
L: Small industry segments may
not be covered
L: None
Case 7:
NDC
Case 2:
2 days of data collection
Case 1
: 3 days of data collection
A: No investment
required
A: Relatively low cost
A: Data appropriate for modeling. Low
cost
A: Modeling needs likely met. Meets OD
survey standards. Available backup data
L
: No I-E, E-I, E-E
trips data
L
: No backup data. No way to verify data
soundness
L
: In case of problems the amount of
backup data is minimal
L
: None
Case 0: NDC
Case 1:
ZIP code Business Patterns Data
Case 2:
30% of records
Case 3
: 40% of records
A
: No investment
A
: Low Cost. Contains summaries of all
observations
A
: Accurate geolocation. Could be
expanded to control totals
A
: Accurate geolocation. Could be
expanded to control totals
L: No data to do
these analyses
L: Data at Zip code level. Unable to
geolocate precisely
L: Data by aggregators may have errors
L: Data by aggregators may have errors
Case 0: NDC
Case 1:
25% of
large establishments
Case 2
: 25% of
large buildings
Case 3:
50% of
large establishments
Case 4
: 50% of
large buildings
Case 5
: Cases 1 and
2
Case 6
: Cases 3 and
4
A: No investment
A: Lower cost
A: Lower cost
A: Low cost
A: Low cost
A: Higher coverage
A: Needs likely met
L: No data about
LTGs
L: Gap in coverage.
No data for large
buildings
L: Gap in coverage.
No data for large
establishments
L: Some gaps in
coverage. No data
for large buildings
L: Some gaps in co-
verage. No data for
large establishments
L: Some groups still
not covered
L: None
Case 0: NDC
Case 1:
200 observations per choice
Case 2:
300 observations per choice
Case 3
: 400 observations per choice
A: No investment
A: Data appropriate for modeling
A: Data appropriate for modeling
A: Data will satisfy most modeling needs
L
: No data to do
these analyses
L
: Require careful design. Data validation
not possible
L
: Small data validation set
L
: None
a
Note
Some of the data could be purchased from data aggregators (e.g., Dun and Bradstreet,
InfoUSA), but may not be as accurate as advertised
Note NDC No data collection, A Advantages, L Limitations, I-E Internal-External, E-I External-
Internal, E-E External-External, OD Origin-Destination GPS Global Positioning System
expected to generate and attract the largest proportion of freight trips and cargo;
thus, they should receive special attention during data collection. Non-freight
related sectors are: finance, insurance and real estate; service industries; and, public
administration. These sectors require or produce some supplies and services for
their operations, which in turn generate freight vehicle trips. For a complete
description and modeling of the freight system they should also be studied.
In addition, it is useful to design a modular strategy for the selection of the
sample size for each data category. Estimates of the sample size are based on the
analyses performed by the authors for the New York City metropolitan region. The
cases shown in Table
8
result in a combinatorial number of potential data col-
lection alternatives and are suggested as examples for implementation in other
large urban areas. These alternatives would differ in the extent to which additional
data are collected and that freight demand modeling is used to synthesize the
missing elements. As typical of these situations, the most appropriate alternative
would depend on the objectives the freight demand model is intended to fulfill, and
the technical and financial constraints at the participating agencies.
An advantage of having such a modular set of alternatives is that they could be
put together as part of a staggered investment in research, model development, and
Search WWH ::
Custom Search