AI based Site Fatal flaw Assessment platform with GIS capability for Renewable Projects, Data Center

Scope of Work (SOW): AI-Based Site Risk & Assessment Platform (Fatal Flaw) - Phase IProject Name: Fatal Flaw Geospatial Risk Assessment Platform (Phase I Prototype)Date: November 2025Duration: [To be defined by agreed milestones]1. Project GoalTo build and deploy a minimum viable product (MVP) featuring an integrated geospatial data pipeline (QGIS/PostGIS), a high-performance Python risk scoring API (FastAPI), and a QGIS-based visualization tool capable of performing "fatal flaw" screening for Solar, Wind, BESS, and Data Center sites across the Continental United States (CONUS).2. Technical Stack (Mandatory)The solution must be delivered using the following technologies to ensure scalability and ease of hand-off:ComponentTechnologyPurposeGIS ClientQGIS (Latest Stable)Primary visualization and input definition tool.Geospatial DatabasePostgreSQL / PostGISCentral repository for all constraint layers with spatial indexing.Backend APIPython 3.11+ / FastAPIHigh-performance service for accepting GeoJSON input and returning scores.Geospatial LibrariesGeoPandas, Shapely, GDAL/OGRData processing, geometric analysis, and ETL operations.EnvironmentDocker Compose or CondaGuaranteed reproducible, versioned environment for deployment.3. Phase I Fixed DeliverablesThe successful completion of Phase I is defined by the delivery of three working, integrated components:3.1. Deliverable 1: Geospatial Data Pipeline and PostGIS SchemaQGIS Project File (.qgz): A configured QGIS project linking directly to the PostGIS database.PostGIS Database Schema: A robust, indexed schema designed to hold constraint layers.Initial Data Load (15 Core Layers): Ingestion, normalization, and projection of 15 critical fatal-flaw layers across the following categories (e.g., using TIGER, NWI, FERC, NREL data):Environmental: Wetlands, Protected Areas (PAD-US), Critical Habitat.Grid: High-Voltage Transmission Lines, Substation Locations (distance calculation ready).Land Use/Zoning: Agricultural Preserves, Federal/Tribal Lands, Steep Slope/High Elevation (DEM derived).ETL Scripting: A set of well-documented Python scripts (.py) using GDAL/GeoPandas to automate the refreshing or loading of these core layers into the PostGIS database.3.2. Deliverable 2: Python Risk Scoring APIFastAPI Service: A complete, runnable Python backend service implemented in FastAPI.Primary Endpoint: /api/v1/assess-siteInput: GeoJSON Polygon (representing the site boundary).Processing: The service will perform the following steps:Validate GeoJSON input.Query PostGIS for intersections/distances against all 15 core layers.Calculate the Fatal Flaw Score using a Weighted Linear Combination (WLC) / Multi-Criteria Decision Making (MCDM) model.Note: The initial weights will be defined by the client but the framework must be flexible to support future ML-optimized weights.Output: A structured JSON response including:overall_risk_score (Normalized 0-100).top_risks (Array of up to 5 highest contributing constraint layers and their impact area/value).fatal_flag (Boolean: True if any constraint layer yields an absolute 'Fatal' result, e.g., intersecting a National Park).API Documentation: Full interactive documentation via Swagger/Redoc (FastAPI native).3.3. Deliverable 3: QGIS Integration and ReportingPyQGIS Plugin: A custom QGIS plugin providing a simple user interface (UI) to:Select a feature (site polygon) from a loaded QGIS layer.Trigger a call to the external FastAPI scoring endpoint.Receive the JSON response.Display the overall_risk_score visually (e.g., color-coding the polygon from Red (Fatal) to Green (Low Risk)).PDF Report Generation: A scriptable component within the QGIS plugin that uses the QGIS Print Layout manager to:Export a professional, standardized, multi-page PDF summary for the selected site.The report must include the calculated Risk Score, the Top 5 Risks breakdown, and a locator map.4. Acceptance Criteria & Hand-off4.1. Acceptance CriteriaThe project will be considered accepted upon successful demonstration of the following:The entire solution runs successfully within the provided Docker/Conda environment.The ETL scripts successfully load the 15 core layers into PostGIS.The FastAPI endpoint processes a complex, multi-vertex GeoJSON polygon query in under 500ms.The system correctly identifies and flags five (5) pre-defined "Fatal Flaw" test sites (e.g., a site intersecting a critical habitat layer).The QGIS plugin successfully triggers the API call, visualizes the score on the map, and exports the PDF report with accurate data.4.2. Project Hand-offThe freelancer will deliver all code and artifacts with clear documentation:Code Repository: Complete source code for the FastAPI service and QGIS plugin.Environment Configuration: Dockerfile, docker-compose.yml, or environment.yml (Conda) files.Hand-off Notes: Comprehensive READMEs explaining setup, deployment, database management, and architecture.