Transportation Data, Surveys and Analytics

Introduction and Context

Transportation engineering is fundamentally an evidence-based discipline. Decisions about roadway capacity, public transport, traffic control, parking supply, freight routes, safety interventions and infrastructure investment depend on the quality of the data used to describe existing conditions and predict future demand. A technically sophisticated model cannot compensate for poorly designed surveys, biased samples, inconsistent definitions or undocumented assumptions.

Transportation data are collected for several purposes: describing current movement, diagnosing operational problems, estimating demand, evaluating projects, monitoring policy, and supporting long-term planning. The appropriate survey method depends on the decision being supported. For example, a short-duration classified traffic count may be adequate for signal timing, whereas a seasonal origin–destination study is required for corridor planning. Similarly, a parking inventory records supply, but an occupancy and duration survey is needed to understand parking performance.

This chapter presents an integrated framework for transportation surveys and analytics. It covers survey design, sampling, traffic and travel-time studies, origin–destination and household surveys, parking and axle-load investigations, statistical analysis, GIS, visualization, forecasting, uncertainty and ethical governance.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Define transportation data requirements in relation to a planning or engineering decision.

2. Design traffic, travel-time, origin–destination, parking, axle-load and household surveys.

3. Apply appropriate sampling, quality-control and data-management procedures.

4. Calculate and interpret fundamental traffic, travel-time, parking and statistical measures.

5. Use GIS, visualization and basic forecasting methods to support transportation analysis.

6. Quantify uncertainty, communicate limitations and make evidence-based recommendations.

7. Recognize ethical, privacy, accessibility and governance requirements in transportation data collection.

1. A Framework for Transportation Data

1.1 Data, information and evidence

Data are recorded observations, such as vehicle counts, speeds, trip origins or parking durations. Information is data organized to answer a question. Evidence is information evaluated for quality, relevance, uncertainty and consistency with alternative explanations.

Transportation data may be classified as:

• Primary data: collected specifically for the current study, such as a manual turning-movement count.
• Secondary data: obtained from existing sources, such as census tables, automatic traffic recorder archives or transit smart-card records.
• Observed data: directly measured behaviour or conditions.
• Reported data: information supplied by respondents, such as stated trip purpose.
• Passive data: generated as a by-product of an activity, such as GPS traces or mobile-device records.
• Active data: obtained through a designed survey or experiment.

A data requirement should be expressed as a chain:

decision → indicator → variable → measurement method → sampling design → quality standard.

For example, the decision may be whether a bus-priority lane is justified. Relevant indicators may include bus travel-time reliability, passenger delay and person-throughput. These require variables such as bus running time, dwell time, passenger counts and traffic volumes, collected across representative periods and conditions.

1.2 Data quality dimensions

Important quality dimensions include:

• Validity: whether the measurement represents the intended concept.
• Reliability: whether repeated measurement under comparable conditions gives similar results.
• Accuracy: closeness to the true or accepted value.
• Precision: degree of variation among repeated measurements.
• Completeness: proportion of required records available.
• Timeliness: suitability of the data for the decision period.
• Consistency: compatibility of definitions, units and coding across datasets.
• Representativeness: extent to which the sample reflects the target population.

Accuracy and precision are not identical. A sensor may produce highly consistent readings that are systematically biased. Calibration, field validation and comparison with independent observations are therefore essential.

2. Survey Design and Sampling

2.1 From research question to survey protocol

A professional survey protocol should specify:

1. Purpose and decisions supported.

2. Study area, target population and observation units.

3. Variables, definitions, units and coding rules.

4. Locations, dates, time periods and required sample size.

5. Instruments, staff responsibilities and safety procedures.

6. Pilot-test arrangements and quality-control checks.

7. Data storage, privacy safeguards and metadata requirements.

8. Analysis methods and reporting format.

A target population is the complete set about which conclusions are required—for example, all weekday commuters entering a central business district. A sampling frame is the operational list or set of locations from which the sample is selected. Coverage error occurs when the frame excludes part of the target population, such as households without listed telephone numbers.

2.2 Probability sampling

Probability sampling gives each eligible unit a known, non-zero selection probability. Common designs include:

• Simple random sampling: units are selected randomly from a complete frame.
• Systematic sampling: every kth unit is selected after a random start.
• Stratified sampling: the population is divided into meaningful groups, such as income bands or geographic zones, and samples are selected within each group.
• Cluster sampling: naturally occurring groups, such as blocks or intersections, are selected and units within them are surveyed.
• Multistage sampling: selection occurs in successive stages, such as regions, neighbourhoods, households and persons.

Stratification can improve precision when strata are internally similar and substantively different. Cluster sampling can reduce field cost but usually increases sampling variance because units in the same cluster tend to be similar.

2.3 Sample size for a proportion

For a large population and a simple random sample, an approximate sample size for estimating a population proportion is

\[

n_0=\frac{z^2p(1-p)}{e^2},

\]

where \(n_0\) is the initial sample size, \(z\) is the standard-normal critical value for the desired confidence level, \(p\) is the anticipated proportion, and \(e\) is the maximum tolerable margin of error. If no prior estimate of \(p\) is available, \(p=0.5\) is conservative because it produces the largest variance. The formula assumes independent random observations and is not automatically valid for clustered or highly weighted samples.

For a finite population of size \(N\), a commonly used correction is

\[

n=\frac{n_0}{1+(n_0-1)/N}.

\]

Non-response requires inflation. If the expected response rate is \(r\), the number approached should be approximately \(n/r\), provided non-response does not introduce serious bias.

Worked Example 1: Household sample size

Suppose a household survey requires a 95% confidence level, a margin of error of 5 percentage points, and no prior estimate of the relevant proportion. Taking \(z=1.96\) and \(p=0.5\),

\[

n_0=\frac{1.96^2(0.5)(0.5)}{0.05^2}=384.16.

\]

Thus, at least 385 completed responses are required under the simple-random-sampling assumptions. If the expected response rate is 70%, the number of households approached should be approximately

\[

385/0.70=550.

\]

This calculation does not account for design effects from clustering, weighting, subgroup analysis or finite-population correction.

2.4 Bias, non-response and weighting

Sampling error is random variation caused by observing a sample rather than the entire population. Bias is systematic deviation. Major sources include undercoverage, non-response, interviewer effects, recall error, social desirability, time-of-day selection and instrument malfunction.

Weights may compensate for unequal selection probabilities or differential response. A basic design weight is the reciprocal of the selection probability. However, weighting cannot reliably repair a missing population segment for which no observations exist. Weight construction, trimming and calibration must be documented because extreme weights may increase variance.

3. Traffic Volume and Classification Surveys

3.1 Purpose and measures

Traffic counts support capacity analysis, signal timing, pavement design, environmental assessment and demand modelling. A count should identify:

• location and direction;
• time interval;
• vehicle classes;
• turning movements;
• lane or approach;
• traffic-control state;
• weather, incidents, roadworks and unusual events.

Common measures include hourly volume, peak-hour volume, directional split, vehicle composition and equivalent passenger-car flow. If a count records \(V\) vehicles during an interval of duration \(t\) hours, the hourly flow rate is

\[

q=\frac{V}{t}.

\]

The peak-hour factor is

\[

PHF=\frac{V_{PH}}{4V_{15}},

\]

where \(V_{PH}\) is the total volume during the peak hour and \(V_{15}\) is the highest 15-minute volume within that hour. The factor ranges from 0.25 to 1.00 when the same direction and vehicle definition are used. A low PHF indicates concentration of flow within the peak hour.

Worked Example 2: Peak-hour factor

A peak hour has 1,800 vehicles. The four consecutive 15-minute volumes are 390, 420, 510 and 480 vehicles. The highest 15-minute volume is 510, so

\[

PHF=\frac{1800}{4(510)}=0.882.

\]

The result indicates moderate within-hour concentration. It should not be interpreted as a capacity value; it describes temporal distribution of demand.

3.2 Collection methods

Manual counts are flexible and suitable for turning movements, vehicle classification and unusual conditions, but they are labour-intensive and subject to fatigue and observer disagreement. Automatic methods include pneumatic tubes, inductive loops, radar, video analytics and Bluetooth or Wi-Fi sensing. Each has characteristic errors: tubes may misclassify closely spaced axles, loops can fail under pavement deterioration, and video systems may be affected by occlusion, lighting and weather.

A robust program combines automatic continuous counts with short manual classification counts. Short counts may be adjusted using permanent counter relationships, but the adjustment must account for day-of-week, season and unusual events.

4. Travel-Time, Speed and Reliability Surveys

4.1 Definitions

For a route of length \(L\) and observed travel time \(T\), average travel speed is

\[

v=\frac{L}{T},

\]

provided compatible units are used. If observations are made over equal-distance segments, the appropriate route-average speed is the harmonic mean:

\[

v_h=\frac{n}{\sum_{i=1}^{n}(1/v_i)},

\]

where \(v_i

eq0\) is the speed on segment \(i\). The arithmetic mean is appropriate when averaging speeds observed for equal time periods, but it can overstate the speed of a vehicle traversing equal-distance segments.

Travel-time surveys may use a floating car, automatic number-plate recognition, Bluetooth re-identification, probe vehicles, GPS traces or transit AVL data. The method should match the required spatial and temporal resolution. Privacy-preserving aggregation is especially important for device-based methods.

4.2 Reliability measures

Travel-time reliability describes the predictability of journey time. Measures include standard deviation, percentile travel time, buffer time and planning-time index. If \(T_{50}\) and \(T_{95}\) are the 50th and 95th percentile travel times, buffer time may be expressed as

\[

BT=T_{95}-T_{50},

\]

and the planning-time index as

\[

PTI=\frac{T_{95}}{T_{free-flow}},

\]

where \(T_{free-flow}\) is a defined uncongested reference time. Percentiles should be estimated from a sufficiently large and representative sample; they are not substitutes for documenting incidents, weather and seasonal conditions.

5. Origin–Destination and Household Surveys

5.1 Origin–destination studies

An origin–destination (O–D) survey identifies where trips begin and end, often with purpose, mode, occupancy, time and route. Applications include traffic assignment, cordon planning, bypass evaluation, freight planning and public transport restructuring.

Methods include roadside interviews, licence-plate matching, postal or online questionnaires, telephone surveys, household interviews, transit passenger surveys and anonymized location data. Roadside interviews can produce rich information but introduce safety, delay and non-response concerns. Plate matching estimates movements without interviewing every driver, but it cannot directly provide trip purpose or socioeconomic characteristics.

The study area is commonly divided into traffic analysis zones. Zone boundaries should be stable, sufficiently detailed for the decision and compatible with available socioeconomic data. An O–D matrix has cell \(T_{ij}\), representing trips from origin zone \(i\) to destination zone \(j\). The row sum is produced trips from zone \(i\), and the column sum is attracted trips to zone \(j\). Matrix expansion and balancing should preserve observed information and make assumptions explicit.

5.2 Household travel surveys

Household surveys capture trips that may not be visible at a single road location. A well-designed instrument records household characteristics, persons, vehicles, trip origins and destinations, departure and arrival times, purpose, mode, access and egress, and whether the trip was linked with other activities.

Important design issues include recall period, proxy reporting, unreported short trips, seasonal variation, respondent burden and accessibility. A one-day diary is operationally manageable but may miss infrequent travel. Multi-day diaries improve observation of variability but increase burden. Questions should distinguish a trip from a stop, and a tour from an individual trip.

5.3 Stated and revealed preference

Revealed-preference data describe observed choices; stated-preference surveys ask respondents to choose among hypothetical alternatives. Stated data can evaluate options not yet available, but hypothetical bias and unrealistic attribute combinations must be controlled. Combining both sources can improve behavioural modelling, provided scale differences and sampling are addressed.

6. Parking Surveys

Parking studies must distinguish supply, occupancy, accumulation, duration and turnover.

• Supply: number of legal parking spaces, including restrictions and accessibility spaces.
• Occupancy: spaces occupied at a specified time.
• Accumulation: number occupied over time.
• Duration: length of stay of a vehicle.
• Turnover: number of vehicles using a space during a period.

If \(O\) spaces are occupied and \(S\) spaces are available, occupancy percentage is

\[

Occupancy=100\frac{O}{S}.

\]

For a study period, turnover per space is the number of vehicles using the facility divided by the number of spaces. A parking survey should record illegal parking, loading activity, disabled-space use, price, restrictions, waiting or cruising, and nearby spillover. A single midday inventory can be misleading where demand varies by time, day, land use or special event.

7. Axle-Load and Vehicle-Weight Surveys

Axle-load data are needed for pavement design, freight regulation, bridge assessment and enforcement. Surveys may use weigh-in-motion systems, static weighbridges or portable scales. Essential variables include axle group, axle load, gross vehicle weight, axle spacing, vehicle class, direction, lane, timestamp and calibration status.

Quality assurance requires sensor calibration, verification against static weights, treatment of overloaded or unusual vehicles, and documentation of missing records. Since pavement damage is highly sensitive to heavy axle loads, averages alone can conceal important tail behaviour. Analysis should include distributions, high percentiles, vehicle-class-specific results and seasonal or route differences. Legal thresholds must be taken from the applicable jurisdiction rather than assumed from a generic standard.

8. Statistical Analysis and Uncertainty

8.1 Descriptive statistics

For observations \(x_1,x_2,\ldots,x_n\), the sample mean is

\[

\bar{x}=\frac{1}{n}\sum_{i=1}^{n}x_i.

\]

The sample variance is

\[

s^2=\frac{1}{n-1}\sum_{i=1}^{n}(x_i-\bar{x})^2,

\]

and the sample standard deviation is \(s=\sqrt{s^2}\). The divisor \(n-1\) provides the usual unbiased estimator of population variance under independent random sampling. Travel times and axle loads are often skewed, so medians, percentiles and graphical distributions should accompany means.

The standard error of the mean under simple random sampling is approximately

\[

SE(\bar{x})=\frac{s}{\sqrt{n}}.

\]

A confidence interval based on a normal approximation is

\[

\bar{x}\pm z\,SE(\bar{x}),

\]

when the approximation is justified. For small samples or unknown population variance, a Student’s t critical value is generally preferable. Complex survey weights and clustering require design-based variance estimation rather than the simple formula alone.

8.2 Correlation and regression

Correlation measures association, not causation. A regression model can estimate relationships between an outcome and explanatory variables, but omitted variables, reverse causality, measurement error and selection bias may invalidate causal interpretation. A model should therefore be assessed through residual diagnostics, out-of-sample validation, sensitivity tests and substantive plausibility—not only by a high coefficient of determination.

8.3 Uncertainty propagation

Uncertainty arises from sampling, measurement, non-response, model specification and future conditions. If an output \(y=f(x_1,\ldots,x_k)\) depends on approximately independent inputs with small uncertainties, a first-order approximation is

\[

Var(y)\approx\sum_{j=1}^{k}\left(\frac{\partial f}{\partial x_j}\right)^2Var(x_j).

\]

When inputs are correlated, covariance terms must be included. For nonlinear models or large uncertainty, Monte Carlo simulation is often more appropriate: repeatedly sample plausible inputs, run the model and summarize the distribution of outputs.

9. Forecasting and Demand Analytics

Forecasting should be tied to a defined horizon, scenario and decision. A simple trend model may be written as

\[

Y_t=a+bt+\varepsilon_t,

\]

where \(Y_t\) is the observed measure at time \(t\), \(a\) is the intercept, \(b\) is the estimated trend and \(\varepsilon_t\) is an error term. This model assumes that a linear trend is a reasonable approximation over the forecast period and that errors are appropriately treated. It should not be extrapolated indefinitely or presented as a behavioural model.

Transportation forecasts may use historical growth, elasticity methods, land-use and demographic scenarios, four-step models, activity-based models, discrete-choice models or machine-learning methods. The traditional four-step framework comprises trip generation, trip distribution, mode choice and traffic assignment. Its practical value remains substantial, but it may represent induced demand, non-motorized travel, time-of-day choice and feedback imperfectly.

Machine-learning models can improve prediction where large, high-frequency datasets exist. However, predictive accuracy does not guarantee causal validity or policy transferability. Models should be tested for temporal drift, geographic transfer, subgroup error, explainability and data leakage. Scenario analysis is preferable to a single deterministic forecast when future land use, fuel prices, technology or policy are uncertain.

10. GIS, Visualization and Data Management

Geographic information systems integrate spatial data such as road networks, zones, stops, crashes, population, land use and terrain. A GIS workflow should address coordinate reference systems, positional accuracy, topology, geocoding quality, scale and temporal compatibility. Spatial joins and aggregation can create the modifiable areal unit problem: results may change when zone boundaries or scale change.

Effective visualizations include:

• time-series charts for demand and reliability;
• histograms and boxplots for speeds, travel times and axle loads;
• desire-line maps for O–D flows;
• choropleth maps for normalized rates rather than raw counts;
• diagrams of turning movements and passenger flows;
• dashboards with filters, metadata and uncertainty indicators.

Maps should use appropriate classification, readable legends, accessible colour schemes and clear denominators. A map of crashes should not imply risk unless exposure, such as vehicle-kilometres or population, is considered.

A reproducible data pipeline should retain raw data, cleaned data, code, metadata, version history and a data dictionary. Cleaning rules must be auditable. Deleted records should be flagged rather than silently removed, and all transformations should be traceable.

11. Illustrative Case: Evaluating a Bus-Priority Corridor

Background

A city experiences unreliable bus service on a 6 km radial corridor. The authority is considering a peak-period bus-priority lane but lacks consistent evidence about delay, passenger demand and effects on general traffic.

Problem

Existing studies used different survey dates and definitions. One used arithmetic mean speed, another reported only the worst peak day, and household data excluded informal settlements near the corridor.

Analysis

The project team first defined the decision indicators: bus travel time, 95th-percentile bus travel time, passenger throughput, general-traffic delay, parking displacement and pedestrian access. It established a stratified survey plan covering weekday peak and off-peak periods, Saturday conditions, school terms and non-school periods, with incident logs and weather records.

Automatic vehicle-location data were matched to surveyed bus runs. A manual passenger count was conducted at selected stops, and a parking inventory recorded legal supply, occupancy and loading activity. A household intercept survey was designed with language assistance and accessible response modes. GPS records were aggregated and stripped of direct identifiers before analysis.

Approach and solution

The team calculated route travel times using complete run times rather than averaging segment speeds. Reliability was reported using median and 95th-percentile values. Passenger demand was expanded using documented sampling weights. A microsimulation model tested alternative lane arrangements, but results were reported as scenarios with sensitivity ranges rather than as a single precise prediction.

The preferred design included peak bus priority, relocation of selected loading activity, improved pedestrian crossing control and monitoring after implementation. A before–after evaluation plan specified the same survey locations, periods and definitions.

Outcome

The analysis showed that buses had moderate average delay but substantial reliability problems. The intervention produced the greatest benefit for passengers during incident-prone periods, while general-traffic effects were concentrated at two intersections. The authority adopted the scheme conditionally, with performance thresholds and a review after six months.

Learning points

1. The decision should determine the indicators before data collection begins.

2. Reliability can be more important than average travel time.

3. Comparable definitions are essential in before–after evaluation.

4. Equity analysis must include groups omitted by conventional sampling frames.

5. A model is most useful when its assumptions and uncertainty are visible.

12. Ethics, Governance and Professional Practice

Transportation surveys can expose personal travel patterns, home locations, income, disability, immigration status and vehicle identifiers. Data collection should follow legality, necessity, proportionality, informed consent where applicable, purpose limitation, access control, retention limits and secure destruction. Public release should use aggregation, anonymization or privacy-preserving methods appropriate to re-identification risk.

Ethical practice also requires safe fieldwork. Surveyors should not be placed in live traffic without suitable traffic management, high-visibility clothing, communication procedures and emergency planning. Respondents should be informed about voluntary participation, expected burden, data use and the right to decline.

Analysts must not selectively report favourable time periods, remove inconvenient observations without justification or present model outputs as measured facts. Procurement documents should specify data ownership, algorithmic accountability, cybersecurity, audit rights and performance validation. Accessibility and inclusion should guide language, interface, location and timing choices.

13. Best-Practice Workflow

A defensible transportation data project generally follows this sequence:

1. Define the decision and theory of change.

2. Review existing data and identify gaps.

3. Define concepts, variables and units.

4. Select the survey design and sampling frame.

5. Pilot instruments, procedures and technology.

6. Train staff and document safety controls.

7. Collect data with timestamps, metadata and quality checks.

8. Validate, clean and preserve raw records.

9. Analyse using methods consistent with the sampling design.

10. Quantify uncertainty and conduct sensitivity tests.

11. Interpret results with domain knowledge and stakeholder evidence.

12. Report limitations, recommendations and an evaluation plan.

Challenges and Limitations

Transportation observations are affected by weather, incidents, holidays, construction, special events and behavioural adaptation. Short surveys may not represent seasonal conditions. Digital traces may overrepresent particular demographic or device-owning groups. Automated recognition may produce unequal error rates across lighting, vehicle type or neighbourhood. Models may embed historical inequities or reinforce them through resource allocation.

No single dataset is universally sufficient. Triangulation—comparing independent sources with different error structures—is often more valuable than increasing the size of one biased dataset. Professional judgment remains necessary, but it should be explicit, reviewable and separated from measured results.

Chapter Summary

Transportation planning and engineering require data that are fit for a defined decision. Survey design begins with the decision, indicators, variables, measurement methods and quality standards. Probability sampling, stratification, response-rate adjustment and documented weighting improve representativeness, but cannot eliminate all bias.

Traffic counts support flow and peak-period analysis; travel-time surveys measure speed and reliability; O–D and household surveys reveal movement patterns and behaviour; parking surveys distinguish supply, occupancy, duration and turnover; and axle-load surveys support pavement and freight decisions. Statistical analysis must reflect skewness, sampling design and uncertainty. GIS and visualization improve spatial interpretation, while forecasting and demand models require explicit assumptions, validation and scenario analysis.

High-quality practice combines technical rigor with privacy, safety, accessibility, reproducibility and ethical governance. The strongest recommendations integrate multiple data sources, communicate limitations and include a plan for monitoring outcomes after implementation.

Key Terms

Target population: The complete group about which a study seeks to draw conclusions.

Sampling frame: The operational list or set of units from which a sample is selected.

Probability sampling: Sampling in which each eligible unit has a known, non-zero selection probability.

Stratified sampling: Sampling separately within defined population groups or strata.

Cluster sampling: Sampling groups of units rather than selecting all units independently.

Sampling error: Random difference between a sample estimate and the population value.

Non-response bias: Bias caused when respondents differ systematically from non-respondents.

Peak-hour factor: Ratio of peak-hour volume to four times the highest 15-minute volume in that hour.

Origin–destination matrix: Array describing trips between origin and destination zones.

Travel-time reliability: Predictability or consistency of journey time.

Planning-time index: Ratio of a high-percentile travel time to free-flow travel time.

Parking occupancy: Proportion of available spaces occupied at a specified time.

Parking turnover: Number of vehicles using a parking space during a specified period.

Weigh-in-motion: Technology that estimates axle and vehicle weights while vehicles are moving.

Standard error: Estimated variability of a sample statistic across repeated samples.

Confidence interval: An interval constructed by a stated statistical procedure to express estimation uncertainty.

Geographic information system: A system for storing, analysing and displaying spatially referenced data.

Data dictionary: A structured description of variables, formats, units, codes and meanings.

Triangulation: Comparison of evidence from different methods or sources.

Design effect: Increase or decrease in variance caused by a complex sample design relative to simple random sampling.

Review Questions

Short Answer Questions

1. Distinguish between data, information and evidence in transportation analysis.

2. State four dimensions of transportation data quality.

3. What is the purpose of a sampling frame?

4. Define parking occupancy and parking turnover.

5. Why should axle-load data be analysed by vehicle class and not only by overall average?

Long Answer Questions

1. Explain how a transportation survey should be designed from the initial decision through data reporting.

2. Compare simple random, stratified, cluster and multistage sampling for a household travel survey.

3. Discuss the methods, strengths and limitations of travel-time data collection.

4. Explain how origin–destination and household surveys support transportation planning.

5. Discuss ethical, privacy and governance requirements for GPS, mobile-device and automated-recognition data.

Analytical / Application Questions

1. A traffic count records 2,400 vehicles in a 30-minute period. Calculate the hourly flow rate and explain why it should not automatically be treated as the design-hour volume.

2. Design a survey to evaluate a proposed parking-management scheme in a mixed-use district. Specify sampling periods, variables, quality controls and analysis measures.

3. A bus corridor has an average travel time of 28 minutes, a 95th-percentile time of 42 minutes and a free-flow reference time of 20 minutes. Calculate the planning-time index and explain what additional evidence is required before recommending a bus-priority intervention.

Multiple Choice Questions with Answers

Question 1

Which sampling method divides the population into meaningful groups and samples within each group?

A. Cluster sampling

B. Stratified sampling

C. Convenience sampling

D. Systematic sampling

Correct Answer: B. Stratified sampling

Question 2

A count records 1,200 vehicles during 15 minutes. What is the equivalent hourly flow rate?

A. 300 vehicles/hour

B. 1,200 vehicles/hour

C. 4,800 vehicles/hour

D. 18,000 vehicles/hour

Correct Answer: C. 4,800 vehicles/hour

Question 3

A peak hour has 1,600 vehicles, and its highest 15-minute volume is 500 vehicles. What is the PHF?

A. 0.64

B. 0.80

C. 1.25

D. 1.60

Correct Answer: B. 0.80

Question 4

Which measure is specifically intended to describe travel-time predictability?

A. Vehicle classification

B. Travel-time percentile

C. Roadway gradient

D. Parking supply

Correct Answer: B. Travel-time percentile

Question 5

For equal-distance road segments, which mean is generally appropriate for route-average speed?

A. Arithmetic mean

B. Geometric mean

C. Harmonic mean

D. Mode only

Correct Answer: C. Harmonic mean

Question 6

If 360 of 450 available parking spaces are occupied, the occupancy is:

A. 20%

B. 40%

C. 80%

D. 125%

Correct Answer: C. 80%

Question 7

What does an origin–destination matrix primarily represent?

A. Pavement thickness by road section

B. Trips between origin and destination zones

C. Vehicle speeds by axle position

D. Parking duration by vehicle colour

Correct Answer: B. Trips between origin and destination zones

Question 8

Which issue is a form of coverage error?

A. A respondent changes an answer after reflection

B. A sensor records duplicate timestamps

C. The sampling frame excludes households without listed contact details

D. A surveyor takes a scheduled rest break

Correct Answer: C. The sampling frame excludes households without listed contact details

Question 9

Why are percentiles useful in axle-load analysis?

A. They eliminate the need for calibration

B. They reveal the upper-tail loading conditions that averages may conceal

C. They convert axle loads into vehicle speeds

D. They guarantee compliance with legal limits

Correct Answer: B. They reveal the upper-tail loading conditions that averages may conceal

Question 10

A facility has a median travel time of 30 minutes and a 95th-percentile time of 45 minutes. What is the buffer time using the stated definition?

A. 15 minutes

B. 30 minutes

C. 45 minutes

D. 75 minutes

Correct Answer: A. 15 minutes