Give the comparison a clear baseline
A coloured network map becomes useful when the reader knows what it represents. Establish the time period, network extent, demand and operating conditions before comparing a proposed change. Keep those details close to the figure so that the comparison can be read correctly.
For a closure study, the brief should identify the affected section, the available alternatives and the travel behaviour assumptions. For a demand study, explain which inputs change and which remain fixed. Each is part of the scenario's meaning.
Separate model building from evaluation
A model's relationship with observed data needs to be documented. NZ Transport Agency's Transport model development guidelines address comparisons between observed and modelled outputs, including calibration and validation. The guidelines explicitly distinguish validation from presenting calibration data in another form.
For a project, the evaluation approach should match the intended decision and the model being used. Report the periods, locations and measures tested, along with the places where performance is weaker or evidence is missing.
Test the assumptions that could change the answer
A useful question for the project team is: which reasonable alternative assumption might change our interpretation? In a closure comparison, that could include the amount of diverted demand or its distribution across available routes. Testing an appropriate range helps reveal where further evidence would be useful.
FHWA's guidance on conducting transportation analyses brings experimental design, calibration, validation, sensitivity analysis and documentation together. Its treatment of assumptions is a useful reminder to retain the decisions made during analysis, rather than leaving them outside the final explanation.
Make the output easy to challenge
We favour a comparison pack that includes the baseline, the proposed change, a clearly labelled difference view and a short account of what the model does not represent. A reader should be able to trace a prominent result back to its inputs and method. Label units and the direction of change consistently. If two maps use different colour scales, explain that difference so that a visual contrast does not mislead the reader.
RoadInsite's road-closure demonstration illustrates this presentation approach using simplified conditions. A project-specific recommendation needs the additional data, modelling and engineering review appropriate to the network and the decision.