Spatial join sanity

Skill buildmoonshot/skillpacks/skills/gis/intermediate/spatial-join-sanity

A beginner-to-expert curriculum of drop-in skills for Claude Code, Codex, and any coding agent. Copy-paste ready, tested, not a link farm.

Install

npx -y skills add buildmoonshot/skillpacks --skill spatial-join-sanity

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when performing a spatial join or location-based filter between two layers — joining attributes by location, or selecting features that intersect, contain, or fall within another layer. Makes the agent pick the correct spatial predicate, confirm CRS and geometry preconditions, and handle one-to-many matches deliberately instead of silently inflating or dropping records.

SKILL.md

2.1 KB, as published. Nobody here has run it

Spatial Join Sanity

A spatial join is where CRS bugs, geometry bugs, and cardinality bugs all converge. Do it deliberately.

Before the join

Confirm both layers share a CRS (see crs-discipline) and have valid geometry (see validate-geometry). A spatial join across mismatched CRSs returns wrong matches with no error.
Pick the predicate that matches the question. Don't default to "intersects" for everything:
- points-in-polygons → within / contains
- any overlap → intersects
- shared boundary / adjacency → touches
- "near, within N meters" → a distance predicate (ST_DWithin, near) — not intersects, and not a post-hoc distance filter.

Handle cardinality on purpose

A spatial join is one-to-many. A point on a shared border matches both polygons; overlapping polygons multiply rows. Decide explicitly:

keep all matches, or
keep the largest-overlap / nearest / first match.

Don't let the join silently change your record count.

Sanity-check the result

Compare input vs output row counts and state it. If a 10,000-point layer becomes 14,000 rows after a join to polygons, that's a boundary/overlap multiplication that needs an explanation — not something to hand over unnoticed. A point count that drops means unmatched features fell out of an inner join; decide whether that's intended.

Why this matters

Spatial joins produce a table that looks normal whether or not it's correct. The errors — wrong predicate, CRS mismatch, silent row multiplication — don't surface as crashes; they surface as a deliverable with the wrong counts. A predicate chosen on purpose and a row-count check catch them before they ship.