Project aim

Tell a developer within half a second whether a candidate site has a deal-breaking constraint, and how risky it is overall, before any money is spent on detailed studies.

  • 15Constraint layers
  • 4Project types screened
  • <500 msAssessment design target
  • PDF + QGISDelivery
Illustrative map for Fatal Flaw Geospatial Risk Platform Illustrative graphic

Overview

Fatal Flaw is a FastAPI service that screens a proposed Solar, Wind, battery-storage (BESS) or data-centre site against 15 authoritative constraint layers such as wetlands and protected areas. It flags any fatal constraint, computes a weighted risk score across the rest, and produces a PDF report. Analysts can run it directly from QGIS through a dedicated plugin.

The challenge

Early-stage site selection usually means opening many datasets by hand and checking each one against a boundary. That is slow, inconsistent between analysts and easy to get wrong. The platform needed to be fast enough for interactive use, explainable enough for a report, and reachable from the GIS tools analysts already work in.

Data & inputs

  • 15 authoritative constraint layers
  • Wetlands
  • Protected areas
  • Candidate site geometries

Approach

  1. 01

    Constraint library

    Fifteen authoritative constraint layers, such as wetlands and protected areas, are loaded into PostGIS and indexed for fast intersection.

  2. 02

    Fatal-flag detection

    Any intersection with a layer marked as fatal for the chosen technology raises a fatal flag immediately.

  3. 03

    Weighted risk scoring

    The remaining constraints contribute to a weighted risk score for the chosen project type.

  4. 04

    Speed

    Redis and spatially indexed PostGIS layers support a sub-500 ms design target per assessment.

  5. 05

    Reports & QGIS client

    Results are delivered as PDF reports and through a QGIS plugin, so screening fits into existing GIS workflows.

Results

Each screening returns a fatal / no-fatal decision, a weighted risk score and a PDF report, from a single API call or from inside QGIS.

Validation

The service is containerised with Docker, and its latency is designed around a sub-500 ms budget per assessment.