ArcGIS Pro and AHP Multi-Criteria Analysis: Practical Guide

Choosing the optimal location for a waste collection centre, a wind farm or a new public facility is never a binary decision. It is a compromise between conflicting constraints: distance to dwellings, road accessibility, terrain slope, regulatory requirements, environmental sensitivity. The methodological tool that structures this kind of trade-off is called spatial multi-criteria analysis(Multi-Criteria Decision Analysis applied to GIS, or MCDA).

This article summarises the fundamentals of Thomas Saaty's AHP method, its implementation in ArcGIS Pro and its QGIS equivalent, and presents an optimal-site study I carried out as part of the Master 1 Geomatics G2M programme.

What is spatial multi-criteria analysis?

Spatial multi-criteria analysis combines several layers of geographic information to produce a suitability map: each pixel receives an aggregated score that reflects how well that location satisfies the full set of selected criteria. It is the standard tool for site selection, zoning, prioritisation of risk areas and vulnerability assessment.

The methodological challenge boils down to two questions. How do we normalise criteria that are not measured in the same units (a distance in metres, a slope in percent, a land-use type as a category)? And how do we weight each criterion to reflect its relative importance in the final decision?

The AHP method in practice (Analytic Hierarchy Process)

Developed by Thomas Saaty in the 1970s, AHP is the most widely adopted weighting method in spatial multi-criteria analysis. It rests on three steps.

First, a pairwise comparison matrix is built: for each pair of criteria, the expert (or a panel of experts) assigns a score from 1 to 9 reflecting the relative importance of the first criterion against the second. The diagonal equals 1, and the matrix is reciprocal (if A is three times more important than B, then B scores 1/3 against A).

Next, the principal eigenvector of this matrix is extracted: its components, normalised to sum to 1, give the relative weights of each criterion. A criterion with a weight of 0.35 will count for 35 per cent of the final score.

Finally, the consistency index (Consistency Ratio, CR) is computed. If CR < 0.10, the matrix is deemed consistent and the weights are valid. If CR exceeds 0.10, the expert has contradicted themselves in their scoring and the matrix must be revised. This statistical safeguard is what distinguishes AHP from eyeballed weighting. These concepts are put into practice in my spatial multi-criteria analysis case study conducted in the first semester of the G2M master's programme, where a six-criterion AHP made it possible to rank candidate sites for a development project.

Once the weights are obtained, two aggregation families coexist. The weighted overlay (weighted sum) remains the dominant method for its readability. Fuzzy logic (fuzzy overlay) complements it where thresholds are gradual rather than sharp (for example, a 14 per cent slope is not categorically different from a 16 per cent slope).

Use cases for local authorities and engineering consultancies

Spatial multi-criteria analysis applies to a broad range of decisions:

• Site selection: wind turbines, photovoltaic plants, waste collection centres, public facilities, urban parks.
• Local urban plan zoning: identification of areas to urbanise, areas to protect, sectors to reclaim.
• Environmental vulnerability: ranking of areas exposed to floods, urban heat islands or soil erosion.
• Urban planning: prioritising neighbourhoods for a planting plan, a cycle network or a depaving strategy.

ArcGIS Pro or QGIS for this type of analysis?

Under ArcGIS Pro, the classic chain mobilises the Spatial Analyst extension (Reclassify tools for normalisation, Weighted Overlay and Fuzzy Overlay for aggregation), ModelBuilder to chain these steps reproducibly, and ArcPy to script the full procedure. This is the reference stack in administrations and consultancies already equipped with ESRI.

Under QGIS, the same chain is fully available through the Processing Toolbox (raster reclassification algorithms, raster calculator, graphical modeller), supplemented by GRASS r.mapcalc for the heaviest operations. The cartographic outcome is strictly equivalent, with zero licensing cost.

My stance in client work: I favour QGIS when the client is not equipped with ESRI (saving an annual licence, deliverable replayable without commercial dependency). I work in ArcGIS Pro when the client is already equipped and the in-house officers will need to maintain or replay the analysis. The AHP methodological expertise stays the same; only the grammar of the tools changes.

The standard deliverables of such a study are three in number: a high-resolution synthesis map PDF for public meetings or reports, a GeoTIFF suitability raster usable in the client's GIS, and a reproducible methodological report documenting the AHP matrix, the computed CR and the normalisation choices.

Order-of-magnitude pricing on CODRUM assignments: cartographic audit from 800 €, full GIS study including AHP, maps and report from 2 500 €. The final price depends on the geographic scope, the number of data sets and the level of deliverables expected: personalised quote within 48 hours after a free initial scoping call. For a specific project, the simplest path is to contact me for a free initial scoping discussion, or to consult my About page for the full background.

For a full breakdown of every service and price range, see the Services & Pricing page.