Planning guide

Planning data analysis use cases: six questions before choosing a method

Six planning questions show how site, place, policy and programme analysis need different geographies, periods, measures and evidence limits.

Method selection examples for NSW planning work. They do not predict approval, feasibility, causation or investment return.

Direct answer

Planning data analysis starts by defining the decision, geography, period and measure. A parcel screen, place comparison, policy baseline, development pipeline and serviceability study need different evidence. Reusing one chart or average across all six tasks creates confident but weak conclusions.

The decision fixes the geography, record and method before analysis begins

A four-part setup prevents a convenient dataset or chart from silently changing the planning question.

  1. Define the decision

    State the site, comparison, baseline, pipeline, serviceability or monitoring question.

  2. Fix the frame

    Choose the geography, period, unit, denominator and source vintage before comparing values.

  3. Select the record

    Use the parcel, place measure, policy baseline, application record or service condition the task requires.

  4. State the boundary

    Separate observation from approval, feasibility, prediction, causation or investment conclusions.

1. Screen a development site

Question: which mapped and written controls may affect this land?

Use the resolved parcel, current environmental planning instruments and the relevant incorporated maps. The output is a source-linked list of identified, unresolved and judgement-dependent checks. It is not an approval prediction.

2. Compare places on one defined measure

Question: how do selected NSW areas differ on housing growth, household change, employment or another planning measure?

Use the same geography type, period, unit and denominator. Show missing values and boundary changes. A difference between areas is an observation. It does not identify the cause.

3. Establish a policy baseline

Question: what conditions existed when a plan, target or programme began?

Freeze the policy date and the source data available at that time. Separate targets, forecasts, scenarios and observed outcomes. Later data can update the baseline analysis, but it should not silently replace the original starting point.

4. Inspect a development-application pipeline

Question: what has been lodged or determined in a place and period?

Use application identifiers, status, dates, development type and location. Keep duplicate, amended, withdrawn and missing records visible. A determination does not prove construction, and a lodgement does not measure future supply.

5. Test land or infrastructure serviceability

Question: how much land meets the stated planning and servicing conditions?

Define every condition and use compatible spatial units. Separate zoned land, serviced land and development-ready land. Those measures answer different questions and should not be merged into one availability figure.

6. Monitor change after a policy decision

Question: did the measured condition move after an intervention?

Use a fixed indicator, cadence and comparison rule. Record other changes that could affect the measure. A before-and-after pattern does not establish that the policy caused it.

Method selection table

Original Plynth evidence

Six-question planning method selector

A Plynth-authored table matching each planning decision to the record it needs and the interpretation error most likely to distort it.

DecisionRequired recordCommon error
Site screenParcel, source, date and proposal factsTreating a map hit as a conclusion
Place comparisonGeography, period, unit and denominatorComparing unlike areas or vintages
Policy baselineCommencement, target and observed measureMixing targets with outcomes
DA pipelineApplication identity, status and timeTreating lodgement as delivery
ServiceabilityExplicit planning and service conditionsTreating zoned land as ready land
Change monitoringStable indicator and confounder noteClaiming causation from timing

Plynth evidence and limits

Plynth publishes bounded NSW research that demonstrates several of these methods. The housing-target intensity study keeps its denominator and geography visible. The industrial-land serviceability study separates zoned, serviced and development-ready land.

Official source starting points

Document every transformation and unresolved data-quality issue.

Change record

What would trigger a recheck

  • A worked dataset, geography, period or official definition changes
  • Plynth changes the analysis outputs represented by the guide
Jurisdiction
New South Wales
Publication status
published
Sources checked
2 September 2026