Map Projection Selection Playbook: Conformal vs Equal-Area vs Equidistant

Date: 2026-03-19
Category: knowledge

Why this matters

Projection choice is one of those decisions that looks cosmetic but silently changes conclusions.

A wrong CRS can make you:

• overstate/understate region-level totals,
• misread accessibility by distance,
• ship visually clean maps with analytically wrong numbers.

USGS’s core reminder is still the right starting point: no flat map preserves direction, distance, area, and shape all at once.

One-line mental model

Pick your projection by the metric you cannot afford to be wrong about.

• Conformal → prioritize local shape/angle
• Equal-area → prioritize area comparison
• Equidistant → prioritize specific distance relationships

Everything else is a controlled compromise.

1) Conformal projections (shape/angle fidelity, locally)

What they preserve

• local angles,
• local shape of small features (not global size).

What they sacrifice

• area and absolute distance (especially away from tuned region/standard lines).

Typical uses

• topographic and engineering maps,
• meteorology and navigation contexts where angular relationships matter,
• regional basemaps where "looks right" geometry is operationally important.

Practical examples

• Mercator (navigation-oriented contexts)
• Transverse Mercator / UTM (regional mapping)
• Lambert Conformal Conic (mid-latitude, east-west regions)

Failure mode to watch

Using conformal maps for area-ranked choropleths (density/resource totals) can mislead quickly at high latitude.

2) Equal-area projections (statistical fairness for regions)

What they preserve

• relative area everywhere (core requirement for area comparisons).

What they sacrifice

• local shape/angles.

Typical uses

• thematic maps comparing quantities by region,
• climate/water/land-use summaries,
• global dashboards where "who is bigger/smaller" must be trustworthy.

Practical examples

• Albers Equal Area (continental/mid-latitude analyses)
• Lambert Azimuthal Equal Area (regional/hemispheric focus)
• Equal Earth (global thematic communication)

Failure mode to watch

Stakeholders complain "countries look stretched." That visual discomfort is expected; metric integrity is the goal here.

3) Equidistant projections (distance truth, but only where defined)

What they preserve

• distance from a chosen point, or along chosen lines (projection-dependent).

What they sacrifice

• area and shape globally (often strongly away from the preserved locus).

Typical uses

• maps centered on one hub (radio range, logistics center reach),
• polar distance analysis from pole-centered maps,
• communication maps where one-to-many distance is the key question.

Practical examples

• Azimuthal Equidistant (distance + direction from center point)
• Equidistant Conic (selected parallels/meridians)

Failure mode to watch

Assuming "equidistant" means all pairwise distances are true. It never does.

4) A fast decision workflow (90-second version)

1. Define the decision (not "make a map," but "compare area", "measure access", etc.)
2. Identify non-negotiable metric (area vs angle/shape vs distance-from-X)
3. Choose projection family (equal-area / conformal / equidistant)
4. Tune by extent (global, continental, local; poles vs tropics)
5. Write projection rationale in metadata (analysis_crs, reason, extent)
6. Render separately in display CRS if needed (often Web Mercator)

If steps 1–2 are skipped, projection selection becomes aesthetic guesswork.

5) Task → projection default table

• Global thematic comparison (totals, rates by country): equal-area
• Local cadastral/topographic editing: conformal (often national grid/UTM)
• "Distance from one center" story/map: azimuthal equidistant centered at that hub
• Web app basemap UX: Web Mercator for display, but analysis in task CRS
• Polar-focused distance products: polar azimuthal equidistant/equal-area based on metric

6) Operational pattern that prevents expensive mistakes

Use a dual-CRS (sometimes triple-CRS) pipeline:

1. Storage CRS (authoritative geodetic or local authoritative CRS)
2. Analysis CRS (selected by metric)
3. Display CRS (often EPSG:3857 for tiled UX)

And enforce guardrails:

• area jobs fail CI if run in non-equal-area CRS,
• distance jobs must declare distance model (geodesic vs planar + CRS),
• output artifact includes analysis_crs and projection_rationale.

7) Red-flag checklist before publishing

• Are we comparing polygon sizes? → If yes, equal-area required.
• Are we making claims about one-to-many distance from a hub? → Equidistant centered correctly?
• Are we reporting universal distance accuracy? → If yes, check claim (usually false).
• Are we using Web Mercator numbers directly for analytics? → stop and re-run.
• Did we document why this CRS was chosen? → if not, treat output as provisional.

Bottom line

Projection choice is not cartographic decoration; it is measurement policy.

• Use conformal when local geometry behavior matters.
• Use equal-area when comparisons by size/coverage matter.
• Use equidistant when specific distance relationships matter.

Then separate analysis CRS from display CRS so your map can be both beautiful and honest.

References

• USGS FAQ, How are different map projections used?
  https://www.usgs.gov/faqs/how-are-different-map-projections-used
• Snyder, J. P. (USGS), Map Projections: A Working Manual (Professional Paper 1395)
  https://pubs.usgs.gov/publication/pp1395
• ArcGIS Learn, Choose the right projection
  https://learn.arcgis.com/en/projects/choose-the-right-projection/
• PROJ documentation (projection parameters and behavior)
  https://proj.org/en/stable/usage/projections.html