Airport Forecasting
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 1/25
Airport Forecasting
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 2/25
Forecasting Demand
• Essential to have realistic estimates of the
future demand of an airport
• Used for developing the airport master plan
or aviation system plan
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 3/25
Master Plan
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 4/25
Data used to predict future1. Airport service area
2. Origins and destinations of trips
3. Demographics and population growth of area
4. Economic character of area
5. Trends in existing transportation activities for the
movement of people, freight, and mail by various modes
6. Trends in national traffic affecting future development
7. Distance, population, and industrial character of nearbyareas having air service
8. Geographic factors influencing transportation requirements
9. Existence and degree of competition between airlines and
among other modes of traffic
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 5/25
Estimates Needed
1. The volumes and peaking characteristics of passengers,
aircraft, vehicles, freight, express, and mail
2. The number and types of aircraft needed to serve theabove traffic
3. The number of based general aviation aircraft and the
number of movements generated
4. The performance and operating characteristics of ground
access systems
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 6/25
Forecasting by Judgement
• Delphi Method: A panel of experts on
different subjects is assembled and asked a
series of questions and projections whichthey take into account to determine a
forecast
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 7/25
Trend Extrapolation
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
1970 1975 1980 1985 1990 1995 2000
Year PAX
1970 198128
1975 259317
1980 295780
1985 340717
1990 360670
1995
2000
375000 390000
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 8/25
Top-Down ModelExtrapolate 1, given 2, get 3:
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 9/25
Cross Classification Model
• Cross Classification: examines the behavioralcharacteristics of travelers
• Travelers broken down into classifications based
upon these characteristics
• Based on the belief that certain socioeconomic
characteristics influence the inclination for travel
• Market study performed to determine the travel
characteristics of the individual groups
• By knowing the different groups’ travel patterns,
forecasts can be made by projecting the patterns out
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 10/25
Factors
• Income
• Occupation
• Age
• Type and location of residence
• Education
• etc…
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 11/25
Market Study
• Market Study method does NOT require complex
mathematical relationships
• uses simple equations to generate a classificationtable or matrix
• Advantage: allows for discrimination between
discretionary and non-discretionary travelers andthe factors that influence both types
Non-discretionary = business traveler
Discretionary = vacationers
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 12/25
Multiple Regression
• Econometric Modeling: relates measures of
aviation activity to economic and social factors
• Multiple Regression is used to determine the
relationships between dependent variables and
explanatory variables
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 13/25
Explanatory Variables
• Economic growth
• Population growth
• Market factors
• Travel impedance
• Intermodal competition
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 14/25
Regression Equations
• Linear Regression form:
Y = mx + b
• Multiple Regression form:
Yest= ao + a1X1 + a2X2 + a3X3 + … + anXn
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 15/25
Statistical Testing of Models
• Tests performed to determine the validity of
econometric models
• The analyst needs to consider the
reasonableness as well as the statistical
significance of the model
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 16/25
Coeff. of Mult. Determination
• Coefficient of multiple determination, R 2 :
measures the variation in the dependant
variable that is explained by the variation inthe independent variables
• (e.g. R 2 1.0 very good relationship)
• Equation:
R 2 =(Yest - Yavg)
2
(Y - Yavg)2
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 17/25
Coeff. of Mult. Correlation
• Coefficient of multiple correlation, R:
measures the correlation between the
dependent variable and the independentvariables
• (e.g. R 1.0 very close correlation)
• Equation:
R = (R 2)1/2
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 18/25
Standard Error
• Standard error of the estimate: measure of
the dispersion of the data points about the
regression line and is used to establish theconfidence limits
• Equation:
(Y - Yest)2
m - (n+1)[ ]y est =
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 19/25
Equations for Trend Line
ENP POP
Year (thousand) (thousand)
1983 469 250
1984 515 260
1985 638 272
1986 758 274
1987 935 287
1988 996 296
1989 1140 307
1990 1361 317
1991 1479 326
1992 1651 332
y = 134.59x + 253.93
R 2
= 0.9872
0
200
400
600
800
1000
1200
1400
16001800
250 260 272 274 287 296 307 317 326 332
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 20/25
Elasticity
• Elasticity: the percentage change in traffic
for a 1% change in fare or travel time
• In the past, it was important
• Even greater significance today due to a
deregulated industry• fare wars
• spoke and hub system
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 21/25
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 22/25
Elasticity
• < -1, Elastic, people may change trip behavior
• E = 0, Perfectly Inelastic, no effect on trip behavior
• -1 < E < 0, Inelastic, insensitive to price
q
p pq = ( )
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 23/25
Elasticity Example
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 24/25
Calculations
• Tourists:(-4000/2) (7/6000) = -2.33 < -1, Elastic
people may change trip behavior
• Commuters:
(-1000/2)(7/7500) = -0.47 -1 < E < 0, Inelastic
insensitive to price
q
p
pq =
( )
7/28/2019 lecture8-1.ppt
http://slidepdf.com/reader/full/lecture8-1ppt 25/25
THE END