CVEN9405 Urban Transport Planning and Practice
The Complete Exam Bible · T3 2026CVEN9405 Overview
- UNSW Sydney
- Term 3 2026
- Two-hour final
- Undergraduate course
Urban Transport Planning and Practice develops the chain of reasoning behind an applied transport forecast.
- Assessed by Online quizzes, an individual assignment and a final exam
- Core workflow Frame, estimate, validate and interpret each transport model
- Hardest step Keeping the denominator and behavioural assumption visible
- Exam focus Show the method, unit, direction and planning consequence
How CVEN9405 is assessed
| Component | Weight | Format |
|---|---|---|
| Online Quizzes | 15% | Individual. The first quiz covers the first two weeks and includes true/false, multiple-choice, numerical and text questions. |
| Assignment | 35% | Individual assignment aligned with the first three course learning outcomes. Consult the current Moodle brief for the deliverable and deadline. |
| Final Exam | 50% | Individual two-hour final exam. The course page directs students to Inspera; consult the current exam entry for all sitting conditions. |
What CVEN9405 covers
The course moves from evaluating transport decisions and understanding land-use feedback to preparing survey evidence, estimating trip generation, modelling mode choice, and validating forecasts. The captured teaching materials support numerical work in present value, sampling inference, origin-destination tables, cross-classified rates, regression and multinomial logit probabilities.
Cost-Benefit Appraisal
Incremental impacts, discounting, NPV, sensitivity and distributionTransport appraisal compares a proposal with a forward-looking base case and counts only the impacts caused by the choice between them.
- Forward-looking base cases and genuinely incremental impacts
- Standing, transfers and the population whose welfare counts
- Physical indicators before monetary valuation
- Real cash flows, discount rates and present values
- NPV as an efficiency measure rather than a complete verdict
- Scenario, one-way and switching-value sensitivity tests
Land-Use and Transport Feedback
Accessibility, location response, time scales and integrated modellingLand use and transport interact through a feedback cycle. The spatial distribution of people and activities creates trips, networks enable those interactions, service levels determine accessibility, and accessibility influences later location decisions.
- Activities, trips, networks, accessibility and location in one feedback loop
- Short-run travel response versus long-run location response
- Accessibility measures built from opportunities and impedance
- Integrated, linked and one-way model architectures
- Spatial aggregation, environmental effects and distributional limits
- Historical and holdout validation for long-range forecasts
Transport Data and Statistical Inference
Sampling units, standard errors, confidence intervals and hypothesis testsTransport datasets contain people, households, trips, days and zones, and those units cannot be exchanged casually. This chapter begins by identifying the independent observation before using sample size.
- Observation units and dependence among repeated trip records
- Sampling distributions and the central limit result
- Standard deviation versus standard error
- Confidence intervals and repeated-sampling interpretation
- Null and alternative hypotheses for two-tailed tests
- Statistical significance, practical magnitude and survey bias
OD Matrices and Cross-Classification
Flow tables, linked records, category rates and denominator checksAn origin-destination matrix counts movement between zone pairs, while cross-classification calculates different trip rates for groups of people or households. Both methods depend on disciplined data handling.
- OD cell counts, row productions and column attractions
- Zone keys, selected-zone subsets and missing-code audits
- Intrazonal trips and the influence of zone resolution
- Person-trip joins and uniqueness on the person table
- One-way and multi-way cross-classification rates
- Survey weights, sparse cells and total reconciliation
Trip Generation Models
Trip rates, category analysis, regression and forecast transferTrip generation estimates productions or attractions before distribution and mode choice. The course material presents trip-rate, cross-classification and linear regression approaches. This chapter compares them as modelling choices rather than a hierarchy.
- Trip generation before distribution and mode split
- Trip-rate models and explicit activity denominators
- Cross-classification as a representation of heterogeneity
- Linear regression predictors, coefficients and residuals
- Observed range, extrapolation and negative predictions
- Forecast transfer under changed household or land-use composition
Random Utility and Mode Choice
Systematic utility, logit probabilities, availability and IIAMode choice models represent an individual's selection among available travel alternatives.
- Choice sets and explicit availability rules
- Random utility as systematic utility plus unobserved error
- Alternative attributes, traveller characteristics and interactions
- Alternative-specific constants and reference alternatives
- Multinomial logit probabilities and relative utility
- IIA, close substitutes and the case for nested structures
Logit Estimation, Validation and Forecasting
Likelihood, model comparison, holdout tests and scenario auditEstimation, validation and forecasting are different claims. Maximum likelihood chooses parameters that make observed choices relatively likely under a specified model.
- Likelihood and log-likelihood for observed choices
- Restricted and unrestricted model comparison
- Likelihood-ratio statistics and degrees of freedom
- Coefficient signs, magnitude and implied value of time
- Independent validation across people, places or periods
- Forecast transfer, behavioural unit tests and audit trails
The course welcome identifies the conventional four-step procedure and places it inside a broader planning system: land use and networks shape data requirements, trip generation, distribution, mode choice and assignment, while validation and acceptance testing determine whether a model should inform a real decision.
The captured teaching material is especially strong on cost-benefit appraisal, land-use transport interaction, sampling and data preparation, trip-generation methods, random utility and discrete mode choice. That combination makes the course both quantitative and interpretive.
You are expected to calculate present values, rates, intervals, predicted trips and choice probabilities, but a number earns its meaning only when its base case, unit, denominator and behavioural assumption are stated. Assessment is split across Online Quizzes worth 15%, an Assignment worth 35%, and a Final Exam worth 50%. The official handbook describes the final as two hours, and the course page directs students to Inspera.
The first quiz notice demonstrates true/false, multiple-choice, numerical and text formats based on the first two weeks. The source capture does not publish a course-specific final section breakdown or permitted-materials rule, so those conditions belong to the current Moodle exam entry. The most reliable study method is to carry one planning scenario through the full workflow.
Define a base case and alternatives, identify incremental impacts, inspect survey units, build an OD or category table, estimate generated trips, model mode shares, and then validate the forecast at the spatial and behavioural scale relevant to the decision. Each step should end with a short audit question: what changed, for whom, in what unit, and which assumption could reverse the conclusion?
This guide uses original calculations and contexts so that you can practise those methods without reproducing proprietary workshop or assessment questions.
Worked example · free
Discount a transport project's net benefits
- +1Write the incremental stream: -3.20 at year zero and +0.95 in years one to four, all in real millions of dollars.
- +1Discount the benefits: 0.95/1.06 + 0.95/1.06² + 0.95/1.06³ + 0.95/1.06⁴ = 3.291 million, allowing rounding.
- +1Subtract the immediate cost: NPV = 3.291 - 3.200 = +0.091 million.
- +1Interpret the narrow positive margin and identify demand, benefit valuation and delivery cost as useful sensitivity tests.
Key terms
- Base Case
- The forward-looking comparison describing what would occur without the proposed option.
- Net Present Value
- The present value of benefits minus the present value of costs across the appraisal horizon.
- Sensitivity Analysis
- A structured test of how uncertain inputs or assumptions affect a model result or decision.
- Accessibility
- The opportunities reachable from a location after accounting for the resistance involved in reaching them.
- Origin-Destination Matrix
- A table whose cells count travel from each origin zone to each destination zone.
- Trip Production
- A trip total attached to the producing end under the model's stated trip-purpose convention.
- Trip Attraction
- A trip total attached to the attracting end under the model's stated trip-purpose convention.
- Cross-Classification
- A method that estimates separate average trip rates for categories defined by selected attributes.
- Standard Error
- The estimated sampling variability of a statistic such as a sample mean.
- Random Utility
- A choice framework that separates observed systematic utility from factors unobserved by the analyst.
- Systematic Utility
- The modelled portion of utility constructed from observed alternative and traveller attributes.
- Multinomial Logit
- A discrete-choice model that converts systematic utilities into probabilities using exponentials and a common denominator.
- Model Validation
- The comparison of model predictions with evidence not used to fit the model.
- Forecast Transfer
- The application of an estimated relationship to a different time, place, population or policy setting.
CVEN9405 FAQ
How is the course assessed?
The published assessment has three weighted components: Online Quizzes worth 15%, an individual Assignment worth 35%, and a Final Exam worth 50%. The numeric rows sum to 100%. Use the current Moodle pages for quiz dates, the assignment deliverable and final sitting instructions.
What is known about the final exam?
The official handbook describes a two-hour final, and the captured course page directs students to Inspera. The available course-specific material does not provide a final section breakdown, question count or permitted-materials rule. Check the current Moodle exam entry and the UNSW timetable for your sitting.
What does the four-step model cover?
The conventional sequence estimates trip generation, trip distribution, mode choice and traffic assignment. This course places that chain within land-use interaction, network and data requirements, validation, acceptance testing and forecasting, so the model is treated as a planning workflow rather than four isolated equations.
How much mathematics and statistics are involved?
Expect applied calculations in discounting, net present value, sampling uncertainty, confidence intervals, trip rates, regression and logit probabilities. The difficult part is usually the setup and interpretation: choosing a denominator, preserving units, checking coefficient direction and explaining whether a forecast can be transferred.
What transport datasets are used in the materials?
The captured workshops use person and trip records to build origin-destination counts and cross-classified trip rates. The important skills are joining on stable identifiers, checking unmatched records, applying the correct observation unit, reconciling totals and documenting whether outputs are raw or expanded.
How should I revise the modelling methods?
Rebuild one small scenario from a blank page. Start with the planning question and base case, prepare a compact data table, calculate the rate or probability, and finish with a limitation tied to the model. Change one input afterward and confirm the output moves in the expected direction.
Questions students ask in each CVEN9405 chapter
- Why is the base case not simply the current situation?
- A base case describes what is expected without the proposal over the appraisal horizon. Demand, maintenance, prices and other policies may change even if the project is not built. Freezing all conditions at today's values can misstate incremental benefits and costs, so each forecast assumption needs a defensible no-project path. Chapter 01 · Cost-Benefit Appraisal
- Why can a transport project change land use?
- A project changes the ease of reaching jobs, services and other opportunities. That accessibility shift can alter the attractiveness of locations to households, firms and developers, subject to prices, land supply and planning controls. The resulting relocation or development then changes travel demand and network conditions. Chapter 02 · Land-Use and Transport Feedback
- Why are twenty thousand trip rows not necessarily a sample of twenty thousand people?
- One person can contribute several trips, making those rows dependent for person-level behaviour. Precision calculations must use the correct observation unit or account for clustering and the survey design. Treating repeated trips as independent people can make standard errors far too small. Chapter 03 · Transport Data and Statistical Inference
- Why do OD matrix row and column totals have different meanings?
- A row collects every destination reached from one origin, so its sum is production from that origin. A column collects every origin feeding one destination, so its sum is attraction to that destination. Both sets must reconcile to the same grand trip total. Chapter 04 · OD Matrices and Cross-Classification
- When is a simple trip-rate model useful?
- A simple rate is useful when its denominator has a clear behavioural meaning, the forecast setting resembles the observations and transparency matters. It becomes fragile when household composition, accessibility, land use or travel costs change in ways the average rate cannot represent. Chapter 05 · Trip Generation Models
- Why does a mode-choice model predict probabilities instead of certain choices?
- Analysts cannot observe every perception, preference or circumstance affecting an individual decision. Random utility treats that missing information as an unobserved component, so people with the same observed attributes are represented through population choice probabilities rather than deterministic certainty. Chapter 06 · Random Utility and Mode Choice
- Why is better in-sample fit not enough for forecasting?
- Adding parameters can improve fit to estimation records while capturing noise or unstable relationships. Forecast use also requires sensible signs and magnitudes, predictor availability, performance on new observations and credible response under policy changes. Fit is evidence, not a transfer guarantee. Chapter 07 · Logit Estimation, Validation and Forecasting
How to study for the exam
Study the course as a connected audit trail. Begin with appraisal because it forces you to name the decision, base case, standing and uncertain inputs. Then move to data: identify whether each row is a person, trip, household or zone, and practise reconciling totals after every join or pivot. For statistical questions, write the estimate, standard error and critical value before substituting.
For trip generation, compare a simple rate, category model and regression on the same scenario so that you see what each method assumes. For mode choice, calculate utilities and exponentials in a table, check that probabilities sum to one, and explain the share movement through the changed attribute.
Keep a direction checklist beside your working: higher travel cost should reduce utility, a higher discount rate should lower future present value, and an unavailable mode should not remain in the logit denominator. End each revision session with a validation question.
Ask what evidence was not used to fit the model, whether the forecast inputs remain inside the observed range, and which spatial or demographic errors could cancel in an aggregate total. In the final weeks, alternate short numerical drills with written diagnoses so you can move quickly between calculation and professional interpretation during the two-hour exam.
