Geospatial engineering
that actually ships.
Spatial databases, custom APIs, analysis pipelines, and interactive maps
built for teams that need location intelligence to be fast, reliable, and usable.
Full-stack geospatial
engineering.
Infrastructure, data, analytics, maps, and automation. All connected, all production-ready.
Cloud & Infrastructure
Spatial infrastructure you don't have to worry about. I design, deploy, and maintain cloud-native geospatial stacks that stay up under real load. PostgreSQL/PostGIS, DuckDB, SQL Server, and ArcGIS Enterprise. Whatever your data lives in, I'll tune it, secure it, and keep it running. Docker on AWS or DigitalOcean, STAC catalogs for imagery assets, PMTiles and Protomaps for tiles without tile-server baggage. CI/CD pipelines so changes ship clean, and IAM so nothing leaks. If it's geospatial infrastructure and it needs to be reliable, that's what I build.
Spatial Data & Analytics
Data doesn't tell you anything until you ask the right questions. I build analytical pipelines that turn raw location data into real answers: site suitability scores, drive-time catchments, territory boundaries, terrain models, and watershed delineations. H3 grids for mobility pattern analysis. Network analysis for routing and service areas. I pull in Overture Maps, OpenStreetMap, NLCD, PAD-US, wetlands, parcel records, and federal datasets so your analysis has the richest possible context. Then I design the geodatabase schemas and ETL to keep it all flowing as your data grows.
Backend & ETL
Spatial data comes from everywhere and feeds into everything. I build the pipelines that move, transform, and serve it. GDAL, Rasterio, and GeoPandas handle the heavy lifting across satellite imagery (Sentinel-2, Landsat, NAIP), elevation models, and bulk vector datasets: NDVI, NDWI, DEMs, COG tiling, all of it. STAC catalogs keep your imagery discoverable. Then it moves into tile serving: GeoServer, pg_tileserv, and Martin for blazing-fast MapLibre vector tiles straight from PostGIS. Custom REST APIs in Node.js, Django, or Flask for anything that needs tighter auth or custom logic. Geocoding, reverse geocoding, and spatial analysis services wired end to end. Even MCP tools so LLMs can talk to your geospatial data.
Frontend & Mapping
A map is only good if someone can actually use it. I build interactive mapping interfaces that people want to work with: MapLibre GL JS, Mapbox GL JS, Leaflet, and Cesium, wired into React or Vue.js applications. Deck.gl for when you need to render millions of data points without the browser catching fire. ESRI Maps SDK for JavaScript when you need ArcGIS enterprise capabilities inside a custom app. 3D terrain, time-dynamic layers, responsive layouts for desktop and mobile. The goal isn't a pretty map. It's a map that helps someone make a faster, better decision.
GeoAI & Automation
I use AI the way a good carpenter uses a power saw: to work faster without skipping the craftsmanship. Claude for architecture design and code reasoning. Codex for implementation speed. OpenClaw for scheduled QA automation and autonomous monitoring pipelines that watch your spatial systems and alert before things break. The result is geospatial software built faster, tested better, and maintained continuously. AI amplifies what I do, it doesn't replace the judgment that comes from designing real systems for real users.
Projects that ship.
Problems that get solved.
Cannabis Real Estate Tool
Property assessment platform combining parcel-level analysis with regulatory compliance data, site suitability scoring, and team collaboration for an industry where location compliance is complex and competitive.
Real Estate Canvassing App
A field intelligence tool for real estate agents who need parcel data wherever they are. Search any property, surface ownership and assessment records, attach notes to specific parcels, save searches for later, and share findings with the team, all from a phone or laptop with no GIS training required. Powered by PostGIS, Node.js, Vue.js, and MapLibre GL.
Inspection Report Automation
Built for Map I.T., a geospatial and land surveying firm delivering precision mapping and GIS solutions. A Node.js application that connects live to ESRI ArcGIS Feature Services, pulls inspection data with site photos from field inspections, then auto-sorts findings into categorized PandaDoc PDF templates. Supports multiple template formats including Short, Long, No Issues, and Project Progress with before/during/after photo grouping. What took hours of manual compilation now finishes in under a minute.
NJ Highlands Sewage vs. Septic Analysis
Spatial analysis for the NJ Highlands Water Protection and Planning Council identifying which parcels across the Highlands Region were connected to municipal sewage systems versus on-site septic. A critical regulatory distinction for watershed planning, development review, and environmental protection. PostGIS queries against parcel records, utility service areas, and environmental constraints, orchestrated with Python, producing the analysis layers planners needed.
Scene Intel
Terrain-aware hiking route planner with least-cost-path routing, Tobler walking-time modeling, and multi-factor cost surfaces. DEM/lidar DTM processing, OSM trail integration, NJ parcel overlay, and real-time scene intelligence, served through a MapLibre GL frontend with Python Flask raster backend and PostGIS.
Raphael Zagha
Founder & CEO at Ziggy Spatial
10+ years of geospatial experience working with government agencies, environmental organizations, and real estate professionals. BA in Environmental Planning with a concentration in GIS from Binghamton University.
Core stack: Python, Django, Flask, React, PostgreSQL/PostGIS, ESRI, MapLibre, Mapbox, Leaflet, GDAL, GeoPandas, Docker, AWS, and modern AI coding/automation tools.
Have a spatial project
worth building?
Tell me about your data, users, and what decision the system needs to support. I'll tell you if I can help and how.