Maps & Spatial Data
Run spatial analysis, generate publication-quality maps, and produce client-ready PDF reports — powered by real data, not AI-generated imagery.
What you can do
- Run spatial analysis — buffers, intersections, overlays, nearest-neighbor, spatial joins
- Generate interactive web maps with popups, legends, layer toggles, and custom basemaps
- Create publication-quality static maps with scale bars, north arrows, titles, and branded colors
- Produce PDF reports combining maps, charts, tables, and written analysis
- Calculate area, distance, density, coverage percentages, and other metrics
- Convert and reproject between formats (Shapefile, GeoJSON, KML, GPX, GeoPackage)
- Clean and merge datasets — standardize CRS, fix geometries, geocode addresses, join attributes
Getting started
- 1
Sign up at claude.ai
Go to claude.ai and create an account with a Pro plan ($20/mo). With a Pro plan, Claude can write and execute Python code directly in the browser — which is what powers the mapping, analysis, and report generation. There's nothing to install.
- 2
Create a Project for your work
On claude.ai, click "Projects" in the sidebar, then "Create Project." Add custom instructions describing your typical deliverables (see "Set up a reusable Project" below). Every conversation you start inside this project will follow those instructions automatically — no copy-pasting prompts each time.
- 3
Start a conversation and upload your data
Inside your project, click "New chat." Use the paperclip icon (📎) to attach files — Shapefiles (zip the folder first), GeoJSON, KML, CSVs with coordinates, Excel files, or GPX tracks. You can attach multiple files at once.
- 4
Describe your analysis
Tell Claude what you need: the analysis, the map style, and the output format. Be specific about projections, symbology, and what the final deliverable should look like. See the example prompts below.
- 5
Iterate, then download
Claude remembers everything in the conversation. Ask for adjustments — change the color ramp, add a scale bar, tweak the buffer distance — without re-uploading your data. When you're happy, click the download links to save your maps, PDFs, data files, or scripts.
Set up a reusable Project
A Claude Project saves your preferences so you don't repeat yourself. Set it up once, then every new conversation inside that project inherits your instructions automatically.
When you create your project, paste something like this into the custom instructions field — adapt it to your own work:
I'm an environmental scientist. When I upload spatial data and ask for analysis: - Use EPSG:4326 unless I specify otherwise - Always include a scale bar, north arrow, and legend on static maps - Use a clean, professional color palette (ColorBrewer) - When I ask for a report, produce a PDF with: title page, maps, summary statistics table, and a brief methods section - Export interactive maps as single-file HTML with popups and layer toggles - At the end of every analysis, save the Python script as a separate downloadable file so I can re-run it later
You can also add files to your Project's knowledge base — reference datasets, report templates, or analysis scripts that Claude produced in a previous conversation. When Claude creates a script as an artifact, save it to the project so it's available in every future conversation without re-uploading.
After this one-time setup, your workflow becomes: open project, start chat, attach new data, describe what you need. Claude already has your preferences, saved scripts, and reference files.
Example prompts
These are meant to be starting points. The more context you give Claude about your data, audience, and deliverable format, the better the result.
Site suitability analysis
“I'm attaching a parcel shapefile, a wetlands GeoJSON, and a roads shapefile. Find all parcels that are: (1) at least 2 acres, (2) more than 200 feet from any wetland, and (3) within 0.5 miles of a major road. Map the qualifying parcels in green on a basemap with the wetlands and roads visible. Include a summary table with parcel ID, acreage, and distance to nearest wetland.”
PDF report with maps and analysis
“Using the attached field survey data (CSV with lat/long, species, and count columns), create a PDF report that includes: (1) a map showing sample locations color-coded by dominant species, (2) a bar chart of species frequency, (3) a summary table of counts per site, and (4) a brief methods section describing the spatial distribution. Format it professionally — I'll be sharing this with a client.”
Comparison map for a presentation
“I have two shapefiles: tree canopy coverage from 2015 and 2023. Create a side-by-side map showing canopy change, and calculate the total canopy loss in acres. Use a diverging color scheme — green for gain, red for loss. Export as a high-resolution PNG suitable for a presentation slide.”
Interactive web map for stakeholders
“Create an interactive HTML map from the attached GeoJSON of proposed trail routes. Color each route by difficulty (easy/moderate/hard from the attributes). Add popups showing trail name, distance, and elevation gain. Include a legend and layer toggles so viewers can show/hide difficulty levels. This will be shared with the public — make it clean and easy to use.”
Reusable analysis workflow
“I need to run a 500-meter buffer analysis around each school in this shapefile, then intersect the buffers with the zoning layer to find how much commercially-zoned land is within walking distance of each school. Give me the results as a CSV and a map. Save the Python script as an artifact so I can add it to this project and re-run it next quarter when the zoning data updates.”
Output formats
Tell Claude what format you need. You can request multiple outputs in one conversation.
- PDF reports — combine maps, charts, tables, and narrative text into a single document. Great for client deliverables, final projects, and stakeholder presentations.
- Interactive HTML maps — zoomable maps with popups, legends, and layer controls. Share as a single file anyone can open in a browser — no GIS software needed on their end.
- Static images (PNG/SVG) — publication-quality maps for reports, slides, and posters. Request specific dimensions and DPI for print.
- Data files — GeoJSON, Shapefile (zipped), GeoPackage, CSV, or KML. Useful when you need to bring Claude's analysis results back into ArcGIS, QGIS, or Google Earth.
- Python scripts — Claude creates these as artifacts you can save to your Project. Re-run them in future conversations with new data, adapt them for similar projects, share them with colleagues, or include them as a methods appendix.
Tips
- Be specific about the deliverable. "Create a PDF report with a map, summary statistics, and a methods section" produces much better results than "analyze this data."
- Mention your CRS. If you know your data's coordinate system, tell Claude. If not, mention where the data came from — Claude can usually infer the correct projection.
- Save scripts to your Project. When Claude produces an analysis script as an artifact, save it to your Project's knowledge base. Next time you start a conversation in that project, Claude already has the script — just upload new data and say "run the analysis." No re-uploading or copy-pasting.
- Build up in one conversation. Start with the core analysis, then layer on: "Now add a legend," "Export that as a PDF with a title page," "Add a chart of the same data." Claude keeps the full context so each request builds on the last.
- Your Project carries forward. Once you set up a Project with your preferences, every new conversation starts ready to go. You won't need to re-explain your output format, CRS, or styling choices — just upload data and describe the analysis.
- Zip your Shapefiles. Shapefiles are multiple files (.shp, .dbf, .prj, .shx). Zip the folder before uploading so Claude gets all the components.