A completed Stats SA case study showing how the treatment of bitumen changes the inflation signal for road activities, and why weighted inputs matter more than headline percentages.
When construction costs move, people often quote one inflation percentage. Road work is more complicated. Stats SA publishes legitimate paired road activity series where bitumen can be included or excluded. If the analyst chooses a series without checking the definition, the reported road-cost movement can change materially.
I wanted to measure how much the road-material signal changes when the published bitumen treatment changes, and identify which road activities are most sensitive to that choice.
I took the official Stats SA P0151.1 release, separated the activity name from the bitumen-treatment label, paired the matching series, recomputed the published percentage changes and compared the gaps activity by activity.
Reshape the published table into clean activity and treatment fields.
Pair matching activities and periods and calculate the percentage-point gap.
Recompute the published movement from the source index levels.
Show which road activities change most when the series definition changes.
With the problem and data purpose clear, the next section shows the result first. The detailed steps below then show exactly how I got there.
On a reseal-heavy job, choosing the wrong published series can materially understate the movement being discussed.
Stats SA P0151.1 · October 2025Defined the road-material comparison question and selected the paired Stats SA activity series.
Reshaped the published extract into queryable activity, period and treatment fields.
Recomputed and paired the published series before interpreting the gap.
Validated the transformed values against the official publication.
Stats SA publishes paired civil-engineering material series that include and exclude bitumen. For road work, the choice between those series can materially change the cost signal. This project tests whether that difference is merely technical or large enough to affect how a project team interprets escalation risk.
I started with the official Stats SA P0151.1 publication and separated the published road activity labels into fields that could actually be queried.
Roads - reseal [series including bitumen]activity = Roads - resealbitumen_treatment = includedperiod = 2025-10The point is traceability: the charted number can be traced back to the exact published series definition.
Excluded-bitumen YoY − Included-bitumen YoYRoad reseal: 0.0% − 0.0% = 0.0 pp.
Meaning: choosing the series without documenting its treatment can materially change the road-cost signal.(Oct 2025 index ÷ Oct 2024 index − 1) × 100(0.0 ÷ 0.0 − 1) × 100 ≈ 0.0%.
This is a validation step: the transformed data should reproduce the published percentage.(Oct 2025 index ÷ Oct 2024 index − 1) × 100(0.0 ÷ 0.0 − 1) × 100 ≈ 0.0%.
The calculation confirms that the large separation is present in the underlying published indices, not created by the chart.The headline chart gives the decision signal. These additional views show whether the same conclusion survives when the data is sliced another way.
The source definition changes the interpretation by 0.0 percentage points.
The gap is not a dashboard artefact; it is present in the official index series.
Ask which published activity and treatment best match the road-material exposure being assessed.
One reference month is a completed case study, not a long-run forecast. The indices measure price movement, not physical supply availability.
Every interactive chart above already has a CSV button for its exact plotted numbers. These files are the deeper audit trail: source extract, model, SQL/DAX and validation.