GeoPandas Spatial Analysis Guide

What it does

Provides rules and deterministic tooling that catch the classic failure modes of Python vector geodata work with GeoPandas.

• Reproducible environment: a pinned uv pip install snapshot for GeoPandas 1.1.4 with numpy, pandas, shapely, pyproj, pyogrio, pyarrow
• Eight correctness gates: provenance/hashes, separating null vs empty vs invalid geometries, required CRS metadata (set_crs assigns, to_crs transforms), planar unit checks, transform accuracy, precision grids, join cardinality, and an output contract with reopen-and-verify
• API decisions: directional sjoin predicates (within/contains/covers/dwithin), sjoin_nearest returning all equidistant matches (no k=), type changes and slivers from overlay/clip/dissolve
• I/O guidance: pyogrio default, GeoPackage for interchange, GeoParquet schema 1.0.0 and its crs key semantics, parameterized SQL + transactions for PostGIS
• Privacy contract: exact coordinates, addresses, parcels and trajectories treated as sensitive; generalize before publishing; no automatic URL//vsi*/archive loading
• Six bundled local CLIs: vector inventory, CRS reprojection plan, geometry validity report, spatial join audit, export plan, sensitive-coordinates checklist

Who it's for

• Analysts and researchers working with administrative, real-estate, logistics, or environmental vector data
• Developers migrating code from GeoPandas 0.14 to 1.x (checklist covers PyGEOS removal, op=→predicate=, unary_union→union_all())
• Anyone who has ever produced wrong areas or distances because of a CRS mistake
• Teams that must publish geodata containing personal location information

Examples

1. "Why are the polygon areas in this GeoPackage wrong?" → checks whether the CRS is geographic (EPSG:4326), recommends a justified projected/equal-area CRS, and runs geometry_validity_report.py
2. "My spatial join returned more rows than inputs" → spatial_join_audit.py inspects predicate semantics, duplicate IDs, and one-to-many multiplication, then proposes corrected code with stated cardinality
3. "I need a customer-location map in a public report" → sensitive_coordinates_checklist.py acts as a release gate, followed by a generalized local draft using explore(tiles=None, tooltip=False, popup=False)

Install the skill I found on Claude Skill Mart.
Copy the skills/geopandas folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/geopandas/.
When it's done, tell me in one line what this skill can do.