Selecting comparable sites in a TRICS trip rate query

Every figure in a Transport Assessment descends from the TRICS site selection criteria you set in the first ten minutes of the query. Choose the wrong land use subcategory, or filter down to four sites in a different region, and the junction modelling, the parking justification and the mitigation package all inherit the error. The highway authority's reviewer will go to the query parameters before anything else, so build the selection to be audited rather than to produce a convenient number.

What the trip rate query is trying to represent

A trip rate query answers a narrow question: how many vehicle, pedestrian, cycle and public transport trips does development of this type, at this scale, in this kind of place, with this amount of parking, generate in a given hour.

The database does that by giving you surveys of real sites. So the selection task is to assemble a sample of surveyed sites that resemble the proposal in the respects that drive travel behaviour, and to exclude sites that differ in those respects. Everything that follows is a judgement about relevance, and judgements have to be explained.

Two framing points worth settling with the officer at scoping.

Whether the assessment is about the net change in trips or the gross trips from the proposal. On a redevelopment with an extant lawful use, the net position usually matters and the existing use needs its own trip rate or a survey.

And which periods are assessed. The network peak and the development peak are often different hours, and for some uses the Saturday peak is the critical one.

Land use category and subcategory

Start at the subcategory, not the category. The database distinguishes, within residential, between privately owned houses, privately owned flats, affordable and local authority housing, mixed private and affordable schemes and retirement or sheltered accommodation. Those produce materially different rates, and picking "residential" as a whole produces a blended figure that represents nothing.

The same applies elsewhere. A discount food store behaves differently from a food superstore. A business park differs from a town centre office. Warehousing differs from an industrial estate, and a distribution unit with a high bay and few staff differs from both.

Where the proposal does not fit a subcategory cleanly, two routes exist. Use the nearest subcategory and explain the mismatch and its likely direction of error. Or build the assessment around a first-principles calculation using staff numbers, shift patterns and servicing schedules, with the database used as a sense check. Mixed use schemes usually need the second approach alongside the first, with internalisation of trips argued explicitly rather than assumed.

Region, settlement type and parking

Three filters carry most of the weight.

Region matters because car ownership, public transport provision and journey lengths vary. Filtering to a single region often shrinks the sample too far, so the usual compromise is to exclude Greater London from a non-London query, and to exclude or isolate regions whose travel characteristics are obviously different from the site's.

Settlement type matters more than region in practice. The database lets you filter by location type, from town centre and edge of town centre through suburban area and neighbourhood centre to free standing sites, and by the population within one mile and five miles. Matching the population bands to the site is the single most defensible filter you can apply, because it is a measurable characteristic of the site rather than an administrative boundary.

Parking is the filter people skip. A residential site surveyed with two spaces per dwelling cannot represent a scheme with 0.5 spaces per dwelling, and a retail site with free surface parking cannot represent one relying on a controlled car park. Set a parking range that brackets the proposal and say what it was.

Date range and older surveys

Older surveys carry habits that no longer hold. Car ownership, internet shopping, home working and food delivery have all moved, and a retail survey from fifteen years ago reflects a different retail economy.

Most authorities now expect a stated date window, and many have a cut-off in their own guidance. Set the window deliberately and say why. Then look at what the window does to the sample: if restricting to recent surveys leaves five sites in the wrong region, the honest answer is to widen the date range and explain the trade-off rather than to widen the geography silently.

Surveys carried out during periods of pandemic restriction are flagged in the database and need a decision. Including them tends to depress trip rates for employment and retail uses and can raise them for residential. The defensible position is to exclude them unless there is a reason to do otherwise, and to say so in the report.

Sample size and the spread it supports

A trip rate from three sites is an anecdote. The database's own good practice guidance is concerned with transparency and with resisting the temptation to narrow a sample until it produces a helpful answer, and reviewers apply the same test.

Look at the spread as well as the mean. The output gives you the individual site rates, and where the highest site rate is several times the lowest, the mean is doing a lot of concealing. That usually means the sample contains sites that are not comparable, and the fix is a better filter rather than a wider confidence statement.

Where the sample is small and the spread is wide, consider:

1. Reporting both the mean and the higher end of the range, and testing the junction capacity against the higher figure.
2. Commissioning a survey of one or two closely comparable local sites, which is quicker and less costly than people assume.
3. Agreeing a sensitivity test with the highway authority rather than arguing for a single number.

The approach to picking the final figure from the sample, as opposed to assembling the sample, is covered in the note on selecting trip rates inside a Transport Assessment.

Recording the query so a reviewer can repeat it

The database produces a report that lists every selected site, the filters applied and the parameters of the calculation. Submit it in full as an appendix. A Transport Assessment that quotes a trip rate without the query output is asking the reviewer to take it on trust, and reviewers do not.

Alongside the output, include a short note in the main text covering the subcategory chosen and why, each filter and its value, the date window and its justification, the number of surveyed sites in the final sample, and any sites excluded manually with the reason for each exclusion. Manual exclusions attract the most scrutiny, so give a reason that relates to comparability rather than to the resulting number.

For a smaller scheme using a Transport Statement, the same discipline applies at a smaller scale. Two paragraphs and the query output are enough.

Where an authority will insist on different filters

Expect the following, and raise them at scoping rather than in a revision.

London boroughs and Transport for London work to their own guidance, generally expecting sites matched on public transport accessibility, trip rates cross-checked against household survey data, and car-free or car-capped schemes assessed accordingly.

County highway authorities often require regional filtering, sometimes to their own region alone, and may publish their own preferred date window.

Authorities with parking stress problems will want the parking filter tightened and will usually ask for a Parking Survey alongside the trip rates, because the trip rate says nothing about where the cars stand overnight.

And where the proposal is unusual, some authorities will reject database rates entirely and ask for a first-principles assessment. That is not an unreasonable position for a use with ten surveyed sites nationally.

Illustrative example: a 60-dwelling scheme of houses and flats in a suburban area of a northern city, 0.9 spaces per dwelling. The query uses privately owned houses and privately owned flats as separate subcategories, excludes Greater London, filters on suburban location type, bands population within five miles around the site's figure, sets parking between 0.6 and 1.3 spaces per dwelling and a date window excluding restriction-period surveys. The sample holds eleven house sites and nine flat sites, the spread on the house rates is modest, and the capacity work is tested against both the mean and the upper end.

PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. For specialist technical reports, qualified professional input may be required, and trip generation work should be prepared or checked by a transport planner. We do not guarantee planning approval. To see the transport documents a scheme is likely to need, enter the site postcode on PlanScope.