Module Title: Transport Data Collection

and Quantitative Research Methods

Your name + Background
CONTENT PART 2
PART 1 • Probability: Introduction and concepts, Probability
distributions functions such as Binomial, Poisson,
Geometric, Normal, Gamma and Beta;
• Sources of transport data • Statistical Inference: Sampling; Point estimates;
• b) transport data in developing countries Confidence intervals; Testing of hypotheses;
Regression and correlation; Analysis of variance
• c) inventory and condition data (ANOVA).
• d) system usage data (demand) • Foundations : Language, philosophy, and ethics in
research; Phases of Management Inquiry
• e) speed, delay and congestion data
• Sampling: External validity; Terminology; Statistical
• f) accident data terms in sampling; Probability and non-probability
sampling.
• g) environmental data
• Survey Research: Types of survey; Selecting the
• h) trip matrix data survey method; Constructing the survey – type of
• questions, question content, response format, question
I) interviews and questionnaires wording and placement
• Report Writing: Key elements; Formatting; Sample
Paper.
Road map
Introduction (Data collection)
• What is data?

Data is defined as a value or set of values representing a specific

concept or concepts.
Data become 'information' when analyzed and possibly combined with
other data in order to extract meaning, and to provide context.
Introduction (Data collection)
In the transportation field data can be collected for the following
purposes:

• Operation planning
• Traffic management
• Perception
Introduction (Data collection)
Operation planning

Data-collection techniques

PUBLIC TRANSPORT

Transit-data-collection techniques required for operations planning can be

divided into three categories:

(i) manual-based methods,

(ii) automated-based methods, and
(iii) AVL-based methods (AVL automated vehicle location).
Data-collection techniques

PUBLIC TRANSPORT
Point check (COUNT)
• Point check is usually described as counts and measurements performed by
a checker stationed at a transit stop.

• The stop selected is virtually the maximum (peak) load point, at which the
transit vehicle departing this stop has, on average, the maximum on-board
load across all route segments.
• A route segment is defined as a section of the route between two adjacent
stops.
• For each vehicle passing the stop, the point check usually contains load
counts, arrival and departure times, and vehicle and route identifications.
Data-collection techniques

PUBLIC TRANSPORT

Ride check
• Ride check refers to counts and measurements performed by either a checker
riding the transit vehicle along the entire route or an automated instrument
(hence, replacing the human checker).
• The ride check contains mainly on and off passenger counts, from which one
can derive the on-board passenger load for each route segment, arrival and
departure times for each stop, and sometimes item-specific surveys or
measurements (vehicle running speed, boarding by fare category, gender of
passengers and baggage size), and record farebox readings.
 Passenger survey (PERCEPTION)

• Transit passenger surveys are conducted in essence while directly confronting the passengers.
• The known survey methods of this type are on-board, at stop, at terminal.
• The most common is the on-board survey. All the surveys are carried out by agency checkers
or drivers or by especially trained staff who either distribute forms to fill out or ask questions
in person.
• One way to increase the response incentive is to hand out a symbolic gift; e.g. a good pen or
key-holder. Such a token may open up more opportunities for cooperation as the gesture is
appreciated.
• In a general survey, multi-type information is obtained, such as O-D, access and egress modes
and distances, trip purpose, routes selected on a trip, fare paid, type of payment, frequency of
use by time of day, and socio-economic and attitude elements.

• Special purpose surveys aim at eliciting only one type of information, such as O-D, opinion of
service changes, transfer activities, pass-holder usage, attitude toward possible fare changes,
or the proportion of different fare types (adult, student, free passes, transfer and special fare).
Population survey
• Generally the population surveys are conducted at the regional level at
home, shop or workplace.

• These surveys are usually interview-based, involving both transit users

and nonusers, in order to capture public attitudes and opinions about
transit service changes (including the impact on household location
decisions), fare changes, and transportation, traffic and land-use projects.

• The population interviews of users and non-users also address the vital
issues of potential ridership, market segmentation, market opportunities
and suggestions for new transit initiatives.
Population survey
Population survey
Population survey
Population survey
Population survey
Data requirements
There is a common thread of data needs across all transit agencies.
• This thread exists as long as agencies share these objectives:
• (a) improving service and operations,
• (b) improving productivity and efficiency by better matching supply and demand,
• (c) improving levels of service through increased reliability as a result of better
control and response, and
• (d) reducing data-gathering, processing, and reporting costs.

Broadly speaking, the data are useful for one or more aspects of route and service
design, scheduling, information system, marketing, deficit allocation, monitoring
management and external reporting.
Data requirements
SOURCES OF TRANSPORT DATA

Mobility Transport Data Overview

In the context of today’s mobility and transport solutions, three
main stakeholders can be identified:

• Public Administrations (PAs),

• Mobility Operators (MOs) and
• final users (FUs).
Public Administrations (PAs)

PAs have a fundamental responsibility for the handling, monitoring,

regulation, planning and management of the city transport systems. PAs
should carry out activities to monitor the mobility infrastructure,
composed of the road network, including streets, cycling paths,
pedestrian pathways, parking facilities, gates, etc., to assess the
infrastructure status and organize possible maintenance operations, as
well as to plan any development to respond to novel/changed needs and
operative needs.

.
Mobility Operators (MOs)

The MOs are public and/or private companies that should offer efficient
and affordable transport services. They can be involved in several
transport modalities; for example, bus, taxi, car renting, car/bike and
scooter sharing/pooling, ride hailing (e.g., Uber and Lyft), etc. MOs’ main
responsibilities include vehicle management, route planning and/or
reservation scheduling, accounting and billing in the agreement with PAs
and other stakeholders.
Final users (FUs)
FUs are city users, persons using the transport service (they are the
demand of mobility): they can be classified as citizens, commuters,
tourists, students, city operators or more generally any individual that
needs to move from one place to another. Their feedback on the quality of
a service is a valuable piece of information to signal problems and
inefficiencies and to identify the priority to be addressed by the PAs and
the MOs to improve the transportation services offered to them
In Figure 1, a schematic
representation of the most
relevant data-kind groups
involved in the mobility and
transport domain is presented.

Figure 1. Schematic representation of the most relevant data-kind groups and derived higher-level information for
smart mobility.
The Organization Data group

Mainly contributes to defining the city infrastructure representation. This

group includes statistical and Geographic Information System (GIS) data
that can provide information on main mobility attractors as positions of
points of interest (POIs) (Residential areas, shopping centers, schools,
entertainment areas, etc.), house and civic number distribution, vehicle
typology, population, road network, positions of traffic signs and lights,
building plant shapes (and with some limitation, even the height of
buildings), raster images depicting cadastral or satellite maps and
descriptions of areas limited to the traffic.
The Organization Data group

Asynchronous information like events produced by traffic officers to

communicate car accidents, street damage, unauthorized processions, etc.,
or temporary regulations issued by PAs to change road directions or ban an
area from all kinds of service or only for some vehicles has to be
considered. This possible information can alter both the infrastructure and
the real-time status of the city mobility environment.

Most of the above-mentioned static data could be classified as open data,

which can be defined as non-privacy-restricted and nonconfidential data
produced and freely released by public or private organizations. A large
part of the open data sets is distributed by public government authorities.
The Sensors Data group

Includes information coming from the network of sensors that can produce
a continuous flow of real-time information. Traffic sensors (spires, virtual
spires, BT sniffer, laser, etc.) are at the basis of any assessment of traffic
conditions, such as counting the number of vehicles traveling on the roads
(and distinguishing among the different kinds of vehicle), estimating
traffic flow, identifying critical conditions and in some cases recognizing
car license plates and thus verifying authorization, insurance, computing
origin–destination matrices, etc.
Other kinds of data are those related to parking sensors, which may
communicate slot occupancy, and in some cases also may perform a match
with the specific car park slot authorization. In most cases, automatic gates
with cameras are used to count entities crossing specific areas, verify plates
and authorizations and issue sanctions/taxations for entrance in restricted
traffic zones and/or entering the city (for example for tourist buses).

Other kinds of sensors can be used to assess weather and

environmental/pollutant conditions; they all may contribute to the
assessment of real-time status concerning the mobility environment since
traffic may produce CO2, NO2 and other specific pollutants.

Additional sensors may be positioned to assess the specific flows of bikes,

pedestrians, public and touristic buses, etc.
The Vehicle and People data,

the third group, are typically produced from

people and vehicles considering them as devices.
They include mobile devices; smart cards used to
obtain access to public transport; on-board units
(OBUs) installed on cars, buses and other
vehicles; mobile cell tower connections; the
geolocalization of smartphones and social media
posts; as well as any usage data from MaaS
services as mobile apps. These data help us
understand the mobility patterns of Fus, travel
behaviour, percepction and can be used to
compute vehicle/people origin–destination
matrices (ODMs).
 2
Zona 1 2 3 4 Total
1
1
2
3
4
3
Total

4
Transport data in developing countries
• Challenges in obtaining mobility data

Infrastructure Deficiencies: Limited transportation infrastructure and a lack

of reliable road networks can hinder data collection efforts. Many areas may
not have adequate physical infrastructure or digital connectivity.

Data Collection Tools: There is often a shortage of modern data collection

tools, including sensors and GPS devices. Traditional methods may be
insufficient for capturing real-time mobility patterns.

Funding and Resources: Many governments and organizations face budget

constraints that limit their ability to invest in data collection and analysis.
Funding for mobility research may not be prioritized.
Technical Expertise: There is often a lack of local expertise in data analytics and
interpretation. Training and capacity-building are essential to improve local
capabilities.

Privacy Concerns: Residents may be wary of sharing personal data due to fears of
misuse or invasion of privacy, leading to resistance to data collection initiatives.

Fragmented Data Systems: There can be a lack of coordination between various

agencies and organizations, resulting in fragmented data systems that make it
difficult to obtain comprehensive mobility data.
High Variability in Mobility Patterns: Mobility behaviors can vary widely across
different regions and communities, complicating efforts to generalize findings or
develop standardized metrics.

Political Instability and Conflict: In some regions, political turmoil and conflict can
disrupt transportation systems and hinder data collection processes.

Cultural and Language Barriers: Differences in languages and cultural contexts can
impact communication and the effectiveness of data collection efforts.

Limited Use of Digital Solutions: In some areas, there is restricted access to

smartphones and the internet, which limits the ability to gather data through digital
platforms.
Case study (topic definition)
• Collect mobility data and compute their travel patterns respondents
and present their trip chains
Correlation
A correlation coefficient is a statistical measurement that shows how
strongly two variables are related. It is a number between -1 and 1, with
higher values indicating a stronger relationship
A positive correlation indicates two variables that tend to move in the
same direction. A negative correlation indicates two variables that tend to
move in opposite directions
Linear regression
Regression analysis is a statistical method that studies the relationship
between variables. It can be used to predict values and identify
important variables
Examples
Binary Logistic Regression is utilized when a researcher desires to
model the relationship between one or more predictors and a
binary dependent variables.

Fundamentally, the researcher is address the question :

What is the probability that a given case fall into one of the two
categories on the dependent variable?
EXAMPLE

In this example we are attempting to

model the likelihood of people
traveling by bicycle in a sample of 898
commuters.

The dependent variable in the model is

cycling_fre. Where 0= no frequent
cycling while 1= frequent cycling

The independent variables (Covariates)

are the set of socio demographic
variables, such as gender, age, income,
education level etc
EXAMPLE
SPSS_Results
EXAMPLE
SPSS_Results
The first step, called Step 0 and includes no predictors and just the intercept.
Often this model result IS NOT INTERESTING TO RESEARCHERS.
Observed- This indicate the number of 0s and 1s that are observed in the
dependent variable.

B- In the coefficient for the constant (also called intersect).

S.E- Standard error around the coefficient for the constant

Wald and Sig- This is the Wald Chi square test that tests the null hypostasis
that the constant equals 0.
This Hypostasis is rejected because the P-value (Sig) is smaller than the
critical P-Value of 0.05. Hence we conclude that the constant is not 0.
Df- This is the degree of Freedom for the Wald Chi-square test.
There is only one degree of freedom because they imply one
predictor in the model, namely the constant.

Exp (B)- This is exponentiation of B coeficient, which is an Odds

ratio.

Score and Sig- This is a score test that is used to predict whether
or not an independent variable would be significant in the model.
In this block the full set of predictors are
entered.

The OMNIBUS test of model coefficients is

the results from the likelihood ration chi-
square tests.

This tests whether a model including the full

set of predictor is a significant improvement in
fit over the null model (Intersect only)

The result shown here indicate that the model

fits the data significantly better than a null
model

X2(6)=45.969; Sig <0.001

Model summary

The model summary table contains the -2 log

likelihood and Two Pseudo – Square measures.
The Pseudo R-Square are interpreted as
analogous to R square (they are not computed
in the same fashion as in the OLS regression).

In general there is no strong guidance in the

literature on how these should be used or
interpreted.

The -2log likelihood (Model deviance) is the

most useful for comparing competing models,
particularly because it is distributed as chi
squares.
Classification Table

Provides the frequency and percentages reflecting the

degrees to which the model correctly and incorrectly
predicts category membership on the dependent
variables

We see

100*129/(129+218)=37.2% of cases that have not

used to cycle frequently were correctly predicted by
the model to not cycle frequently

100*443/(108+443)= 80.4%

The overall classification accuracy based on the

model is 63.7%
 Variables in the equation

B- Regression coefficient. Corresponding the

change in the log odds per unit increase on the
predictor.

EXP (B) Odds ratio is interpreted as the

multiplicated change in odds for every one unit
increase on a predictor

In general an odd ration >1 indicate that as scores

on the predictor increase, there is an increasing
probability of the case falling into the target group
on the dependent variable.

If the OR=1, then this indicates no change in the An odd<1. can be interpreted as decreasing
probability of being in the target group as scores on probability of being in the target group as scores
the predictor change. on the predictor increase.
Gender is positive (0.908) and significant
(Sig=0.000) predictor of the frequent
cycling behaviour, with OR indicating that
every one unit increase on this predictor
(gender), the odds of cycling frequently
change by a factor of 2,480 (meaning the
odds are incensing)
QUIZ 1
USING the data ALL segments, please explain how the different socio
demographic variables, age, gender, income predict cycling Travel time
(tt).

Please explain by presenting tables and short text

Submission
link:
https://drive.google.com/drive/folders/
112h4GPfxTWZP3_4psIJ5oVJSnSiAJoGh?usp=drive_link
Case study
Students are requested to develop a questionnaire, to access how people in Kigali
are satisfied with the quality of Public transport service (Buses)

a) Using SPSS, explore the correlation between different variables explaining the
bus service and socio demographic variables
b) Explore how the different variables explain bus service quality can predict the
number of bus trips (linear regression)
c) Explore the likelihood of people shifting from all the modes to buses (Binary
logistic regression)

Write a report and please use endnote or other app for citation.