Professional Documents
Culture Documents
TP Group7
TP Group7
Group Project
Trip Generation Model for IITR Hostel Students - Aggregate & Disaggregate
Analysis
Introduction
Trip generation is the first and foremost phase in the four stage travel demand modeling
process. It predicts the number of trips originating in or destined for a particular traffic
analysis zone (TAZ). In general, it is influenced by numerous factors which include
socioeconomic characteristics of the trip maker, land use characteristics, development of
the area, distance from the city or urban centers. Development of the area whether it is
an urban or suburban area is one of the major factors that affects the trip generation.
Modeling basically relates the dependent variable i.e. trips produced by a zone for
aggregated model or household trip production rate for household based models to the
corresponding Independent variables characterized by the whole zone or household
characteristic respectively. Calibration is done based on the set of observations obtained
corresponding to the zones for aggregate model and for disaggregate model employs a
number of base year observations corresponding to an individual household in a sample
of households drawn randomly from the region.
Project overview
Problem statement
Explore the trip generation (both production and attraction) model for IITR hostel
students using different methods (aggregate and disaggregate analysis).
Objectives
Task 1: Aggregate Analysis
1.1. Collect Data: Collect data on the number of hostel students, their demographics
(gender, age, etc.), the distance between the hostel and the departments/ markets
1.2. Analyze Data: Analyze the collected data to identify the factors that influence a trip
generation by hostel students. Use statistical techniques such as regression analysis to
estimate the relationship between the number of trips generated and the independent
variables.
1.3. Develop a Model: Develop an aggregate trip generation model that estimates the
total number of trips generated by hostel students based on the identified factors.
1.4. Validate the Model: Validate the model by comparing the predicted number of trips
with the observed number of trips generated by hostel students.
1.5 State the applicability of your model for the new hostel close to BEG gate.
2.2. Analyze Data: Analyze the collected data to identify the factors that influence the
travel behavior of hostel students. Use statistical techniques to estimate the relationship
between the dependent and independent variables.
2.3. Develop a Model: Develop a disaggregate trip generation model that estimates the
number of trips generated by individual hostel students based on their travel behavior
and other factors.
2.4. Validate the Model: Validate the model by comparing the predicted travel behavior
with the observed travel behavior of hostel students.
3.1. Compare the results of the aggregate and disaggregate analysis to identify the
similarities and differences between the two models.
3.2. Interpret the results of the models to identify the factors that influence a trip
generation by hostel students and their travel behavior.
3.3. Discuss the implications of the findings for transportation planning and
management in and around IIT Roorkee campuses.
Site overview
Kasturba Bhawan, the second Girls' Hostel in IIT Roorkee, was established initially as a
two-floor building in the year 2006. As the time passed and the need to house the
increasing number of girl entrants to the institution increased, it was extended four
times and now in its present form is seen as a vibrant seven floored, two winged
multi-coloured building inaugurated in 12 November 2010. The name of the bhawan is
dedicated to Kasturba Ji, who is an inspiration for all girl residents. The Bhawan has
every facility that one needs to live a comfortable life. We live here, not just to survive
but to provide an expression to our lives and this Bhawan is the foundation of it all.
Objective 1 - Aggregate analysis
1.1
Architecture 450 3
Chemistry 950 8
Hydrology 900 7
mathematics 950 8
Physics 950 7
SAC 100 2
MAC 1200 10
Temple 650 5
Library 800 6
Playground 750 6
1.3
ii. Regression Analysis: For the aggregate model, we want to estimate the
relationship between Distance from Kasturba Bhawan (d_in_km), No of students for
each destination (no_of_students), the average income of these students (avg_income)
and the average age of these students (avg_age) with Dependent variable which is the
total number of trips for the destination. To estimate the relationship between a
number of trips generated (Dependent Variable) and independent variables (Distance
from Kasturba Bhawan, No of students for each destination, the average income of
students and the average age of students), we will use a linear regression model. For the
estimation of the results, we have used Statistical Software called “STATA”.
No. of trips generated for destination = f (Distance, No. of students, Average age,
average income)
Regression Equation :
Where b0, b1, b2, b3 and b4 are coefficients and Ui is the error term.
b0 = Intercept term
Results
No. of Observation 32
F(4,27) 10.5
R^2 0.6028
Table : ANOVA
Source ss df MS
As per the above regression analysis, we can identify the independent variables that
have a significant impact on the dependent variable. So we will include
d_in_km(distance), no_of_students, and avg_income(Average Income) and we can
exclude avg_age (average age). The aggregate trip generation model is as follows:
Model is good in predicting the number of trips for those zones, which don't have
extremely high or low values of trips. The cause of this result is the high value of the
intercept term (_cons) in the model. From the table below, we can compare the
predicted and observed trips for different zones.
Hydrology 79 238
1.5 Applicability of the model for the new hostel close to BEG gate -
Since students have homogeneity in their travel patterns and other variables related to
demography etc., we can assume that other factors are the same for the new hostel. We
have predicted the number of trips generated weekly for different zones based on the
distance of different zones from BEG gate while taking other factors same as for the
Kasturba Bhawan. It helps decision-makers to understand the potential impact of the
project and make informed decisions about the allocation of resources. We have
presented the table below which includes the distance from BEG gate (in meters) and
trips predicted (weekly). Predicted trips can help in transport planning and planning for
the infrastructure required for managing the traffic flow. It can help in identifying
potential bottlenecks and the need for additional roads or lanes. This includes
determining the number of lanes, the type of roadway, and the expected flow of traffic.
Earthquake 1100 31
department
Define trip purposes: Define the purpose of trips, such as work, shopping, education,
or recreational activities. This information will help in identifying the types of trips
made by individuals and their trip-making behavior.
Develop models: Develop models to estimate the number of trips made by individuals
based on their socio-economic and demographic characteristics, land-use patterns, and
transportation infrastructure. These models may be regression models or other
statistical models that help in identifying the factors that influence trip-making
behavior.
Analyze data: Analyze the data using the models developed in step 3 to estimate the
number of trips made by individuals for each trip purpose. This will help in
understanding the factors that influence trip-making behavior and how different
socio-economic and demographic characteristics, land-use patterns, and transportation
infrastructure affect trip generation.
Validate the models: Validate the models using different statistical techniques such
as cross-validation, sensitivity analysis, and goodness-of-fit tests. This will help in
ensuring that the models are robust and provide accurate estimates of trip generation.
Use the models: Once validated, the models can be used to forecast trip generation for
different scenarios, such as changes in land-use patterns, transportation infrastructure,
or socio-economic and demographic characteristics. This information can be used by
transportation planners and policymakers to plan for future transportation
infrastructure and to make policy decisions related to land use and transportation
2.1 Data collection
For collecting data on an individual level ,a questionnaire was prepared and circulated
among the Kasturba Bhawan hostel students.
Out of 859 students residing in the hostel,91 responses were recorded through
questionnaire surveys i.e approximately 10.6% of the total residing students.
From the sample taken,about 69% of the students were UG students ,22% of the
students were doing PG and 9% of them were from PhD.
Most of the students were from UG 4th year,followed by UG 3rd year,PhD 2nd year etc.
The least were from PhD 4th year.
Average age of PhD students was 29 years, whereas UG and PG students had an average
age of 20 and 26 respectively.
From the graphical representation of the data,we can see that most of the PhD scholars
use other modes of transport which includes walking,2 wheelers.Only 25% of total Phd
scholars use cycle.Most of the PG students use cycle for commuting whereas UG
students depend on E rickshaws.
Students belonging to the age group of 26+ use other modes of transport like walking ,2
wheelers etc for commuting while the age group of 24-26 are the most common users of
cycle. The 18-20 age group prefer E rickshaws for transportation.
The average daily trip frequency for attending lectures for UG students is found to be 6
while that of PG and PhD is 4.
The average weekly trip frequency for attending classes for UG students is found to be
14 while that of PG and PhD is 13 and 11.
The average weekly trip frequency for non class activity for UG students is found to be
16 while that of PG and PhD is 14 and 18.
PG 12400
PhD 31000
UG 0
The monthly income of students according to the courses are listed in the table
above.Income level is obtained on the basis of the scholarship they receive monthly.
Also we can see that depending on the income,the number of trips made by PhD
scholars for non class activities is more compared to the PG and UG student’s trip
frequency for the same.
The peak hour time period when most of the trips are made are during 9-11 am in
the morning, 12-2 pm at noon and 5-7 pm in the evening.The peak hour during
weekdays and weekends vary. During weekends, peak hours in the evening are between
6-9 pm. 60% of the respondents leave the hostel during 9-11 am in the morning for
lectures.Almost 80% of the respondents return to the hostel between 12-2pm for lunch
and 50 % leaves hostel to attend classes.70% of the trips are generated during the
evening peak hour between 5-7 pm when the respondents come back to hostel after
classes and goes out for social activities,temple,sports,students club,shopping etc.
2.2
I- Factors that influence the travel behavior of hostel students - From the
above data we can see that there are many factors that affect individual travel behavior.
From the data, we have identified a few factors that affect the travel behavior of hostel
students at individual level. These factors include Gender, Age, Travel Mode (Cycle,
E-rickshaw, and others), and Course (UG, PG, PhD), No. of classes daily, Day (Is it
weekend or not), Purpose of the trip.
Regression Analysis - For the aggregate model, we want to estimate the relationship
between dependent variable (total number of trips generated by each individual) and
independent variable (Individual Income Level, Trip frequency for daily classes, Mode
used (Cycle, E-rickshaw or others),age, year, course of study (UG, PG or PhD)). Two
independent variables are categorical in nature 1)- Mode Used 2)- Course of the study.
Mode used has three categories (Cycle, E-rickshaw, Others) and the Course of the study
also has three categories (UG, PG or PhD). To estimate the relationship between the
number of trips generated by each individual (Dependent Variable) and independent
variables, we will use a linear regression model. Since we have a categorical variable in
the model, we have to leave one category of each categorical variable to avoid Dummy
Variable Trap. So we are dropping e-rickshaw from modes of travel and UG from
courses of study. For the estimation of the value of the coefficients, we have used
Statistical Software called “STATA”.
No. of trips generated by each individual = f (Individual Income Level, Trip frequency
for daily classes, cycle , other mode, age, year, PG, PhD)
Regression Equation(For Disaggregate Analysis):
Where b0, b1, b2, b3, b4, b5, b6, b7, b8 are coefficients and Ui is the error term.
b0 = Intercept term
b3 = change in number of trips due to use of other modes except cycle and E-rickshaw.
Results
1- Model Statistics - Model has been estimated for 91 observations (as per the
responses to the questionnaire). Since probability > F = 0.000007 so the model is
significantly explaining the number of trips generated. Adjusted R^2 = 0.6668 is also
good which is a measure of the goodness of fit of the model. R^2 is 0.6964 which means
69.64% variation in the number of trips generated for a particular origin and destination
(dependent variable) is explained by model-independent variables.
Table: Model Statistics
No. of Observation 91
F(8,82) 23.51
R^2 0.6964
Table: ANOVA
Source ss df MS
As per the above regression analysis, we can identify the independent variables that
have a significant impact on the dependent variable. So we will include
individualincome_level(Individual Income), tfclassesdaily(Trip frequency for daily
classes), cycle (mode of travel), age, year (year of the study), coursepg (course of the
study) and we can exclude coursephd and others (mode of travel). The aggregate trip
generation model is as follows:
For the validation of the model, we will predict the total number of trips generated by 10
students on the basis of independent variables that are included in the model. And we
will compare the actual value and predicted weekly total trip frequency for each student.
1 21 22
2 23 23
3 27 26
4 26 27
5 14 19
6 23 24
7 18 24
8 5 26
9 21 18
10 19 22
Objective 3: Comparison and Interpretation
Probability > F = 0.000012 so that model Probability > F = 0.000007 so the model
is significantly explaining the number of is significantly explaining the number of
trips generated. trips generated.
Adjusted R^2 = 0.544 is also good which Adjusted R^2 = 0.6668 is also good
is a measure of goodness of fit of the which is a measure of the goodness of fit
model. of the model.
R^2 is 0.6028 which means 60.28% R^2 is 0.6964 which means 69.64%
variation in the number of trips generated variation in the number of trips generated
for particular origin and destination for a particular origin and destination
(dependent variable) is explained by (dependent variable) is explained by
model independent variables. model independent variables.
The distance from a specific destination affects the no. of trips generated. Average
income and no. of students residing in the bhawan also plays a significant role.
Results from disaggregate analysis gives a better explanation for the no. of trips
generated from Kasturba Bhawan, considering its R^2 value is better than the aggregate
model.
The results may or may not be applicable to other areas in and around IIT Roorkee, as
no. of trips generated would majorly depend on the factors generating the trips, which
may be zone specific or individual specific, depending on the analysis method used. The
number of trips may vary depending on the above factors, which is why the exact same
model won’t necessarily hold true for other bhawans/areas in and around the campus.
As most trips are generated for SAC, if we assume a similar kind of pattern for other
bhawans/hostels as well, considering the mode of transport to be majorly walking or
cycling according to the survey results, we can promote the provision of dedicated
paths/lanes for such users to avoid accidents and mixed traffic.
As most of the trips observed are nearby places only, we can predict that students tend
to traverse around the campus, preferably to amenities available nearby.
Efforts By -