Professional Documents
Culture Documents
1 Transport Data
1 Transport Data
Urban Transportation
Planning and Management
Data collection
Sampling vs population
Sampling approaches
Space
Zoning systems
Networks
Data
Travel is a phenomenon that takes place in time
and space
Where travel will take place
When travel will take place
Travel Data
Disaggregate - Data is always collected at this level
Aggregate - Data is often reported at this level
Space-Time Path
Mei-Po Kwan
3-D GIS gallery
at Ohio State
University
Zoning Systems
Portland, Oregon
Space-Time
Paths, African-
American Women
Mei-Po Kwan
3-D GIS gallery
at Ohio State
University
Space - Networks
Transportation Networks
Supply side of the transportation system:
what the system offers to satisfy movement needs of travelers
Population
Definitions
Sample
Population
Sample
Sampling
Sample size Trade-offs
Accuracy – Small
sample size is cheaper
but may be subject to
unacceptably high
variability
Sampling
In most cases, population is large and it is impractical as
well as costly (time, money) to study all people.
* Step 1:
An identifier (number) is associated
to each unit/individual in the
population
* Step 2:
Selection of numbers randomly to
obtain the sample
! With stratified
$0 – $10,000 R sampling,
$10,001 – $20,000 R minorities can
$20,001 – $30,000
R be identified
and stratified
> $100,000 R
Random
Population Selection Sample
Procedures
Stratified Random Sampling
* Stratify by Family Size:
Family No. of
Size households No. of trips
0 1 2
2
150 250 1121 1500 350 450
3 or
more 180 1500 1002 2000 230 500
Choice-based Sampling
The population is stratified based on the result of a choice process under
consideration (mode choice: car, metro, etc.).
* Main Disadvantage
The sample may not be random and
therefore there might be risk of bias.
2. Sampling Bias
Errors of Sampling
1. Sampling Error
This is just because we are dealing with a sample and not with
the entire population. It is always present due to random
effects.
This type of error does not affect the expected values of the
means of the estimated parameters; it only affects the
variability around them.
population
sample
mean = µ
Errors of Sampling
2. Sampling Bias
It is caused by mistakes while defining the population of
interest, when selecting the sampling method, the data
collection technique (mail, telephone, personal interview)
or any other part of the process.
population
sample
bias
mean = µ
Errors of Sampling
Sampling (Measurement) Error Sampling Bias
population
population
sample
sample
mean = µ
bias
mean = µ
1 1 1
S se se se
se( x ) = 2 4 8
n 4n 16n 64n
Sampling Approaches (Example)
Population