feat(ggplot2): implement map-projections

[github-actions]

Implementation: map-projections - r/ggplot2

Implements the r/ggplot2 version of map-projections.

File: plots/map-projections/implementations/r/ggplot2.R

Parent Issue: #3771

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

Light render (plot-light.png): A Mollweide equal-area projection world map on a warm off-white (#FAF8F1) background. The elliptical projection boundary is filled with light blue ocean (#C4DCF0). A subtle gray graticule (lat/lon grid at 30° intervals, linewidth=0.18, alpha=0.45) covers the ellipse, with the equator and prime meridian rendered slightly bolder (alpha=0.70). Thirty green (#009E73) Tissot indicatrix ellipses are distributed across 5 latitude bands (±60° in 30° steps) at 6 longitude positions (±150° in 60° steps). No land masses or country borders are visible — the entire interior is ocean-colored. Title "map-projections · r · ggplot2 · anyplot.ai" in dark ink at 12pt, centered. Subtitle in gray at 8.5pt explains the equal-area property. All text is clearly readable. Legibility verdict: PASS.

Dark render (plot-dark.png): Same composition on a warm near-black (#1A1A17) background. The ocean switches to dark navy (#152030) — semantically appropriate and visually distinct. The graticule appears as subtle off-white paths. The Tissot indicatrices remain exactly #009E73, identical to the light render — only chrome flips, data colors are unchanged. Title and subtitle use light ink, clearly readable against the near-black surface. No dark-on-dark failures anywhere. Legibility verdict: PASS.

Both renders are fully theme-correct. The only difference between themes is the ocean color and chrome inversion — the green Tissot ellipses are identical in both.

Score: 83/100

Category	Score	Max
Visual Quality	29	30
Design Excellence	15	20
Spec Compliance	13	15
Data Quality	13	15
Code Quality	8	10
Library Mastery	5	10
Total	83	100

Visual Quality (29/30)

• VQ-01: Text Legibility (7/8) — Title (12pt) and subtitle (8.5pt) explicitly set; readable in both themes; subtitle is on the smaller side for mobile scaling
• VQ-02: No Overlap (6/6) — No overlapping elements; Tissot ellipses well-spaced
• VQ-03: Element Visibility (6/6) — Tissot indicatrices clearly visible and well-sized; graticule subtle but legible
• VQ-04: Color Accessibility (2/2) — Green (#009E73) on blue ocean has good contrast in both themes; CVD-safe
• VQ-05: Layout & Canvas (4/4) — Map fills ~75% of canvas, balanced margins, nothing cut off
• VQ-06: Axis Labels & Title (2/2) — Title descriptive; subtitle explains the projection and Tissot purpose; theme_void map needs no axis labels
• VQ-07: Palette Compliance (2/2) — Tissot ellipses use #009E73 (position 1); ocean blue is a strong semantic association; backgrounds correct #FAF8F1/#1A1A17; chrome flips correctly

Design Excellence (15/20)

• DE-01: Aesthetic Sophistication (6/8) — Strong design: custom dual-theme ocean colors, intentional graticule opacity hierarchy, harmonious green/blue composition — clearly above defaults
• DE-02: Visual Refinement (5/6) — theme_void() removes all extraneous chrome; graticule very subtle (alpha=0.45); equator/prime meridian differentiated at alpha=0.70; clean ellipse outline
• DE-03: Data Storytelling (4/6) — Subtitle explicitly explains what Tissot indicatrices demonstrate; visual hierarchy guides eye from graticule to reference lines to Tissot ellipses

Spec Compliance (13/15)

• SC-01: Plot Type (5/5) — Correct: Mollweide projection world map; spec allows single projection with clear title
• SC-02: Required Features (2/4) — Graticule at 30° ✓, Tissot indicatrices ✓; missing country boundaries and coastlines which the spec explicitly requires ("Show clean coastlines and country borders with appropriate styling")
• SC-03: Data Mapping (3/3) — Mollweide math correctly implemented; graticule correctly positioned; Tissot circles at valid lat/lon centers
• SC-04: Title & Legend (3/3) — Title exactly "map-projections · r · ggplot2 · anyplot.ai"; no legend needed for single-series map

Data Quality (13/15)

• DQ-01: Feature Coverage (4/6) — Shows projection math, graticule, Tissot indicatrices — but missing the key world map feature (land polygon fills and country outlines)
• DQ-02: Realistic Context (5/5) — Real geographic context; Mollweide is a well-known standard equal-area projection; educational scenario; neutral content
• DQ-03: Appropriate Scale (4/4) — Projection math correct; Tissot radius (5°) textbook-standard; lat/lon centers appropriate

Code Quality (8/10)

• CQ-01: KISS Structure (2/3) — Has helper functions solve_theta() and mollweide(); not pure flat KISS, though they are mathematically necessary for the projection
• CQ-02: Reproducibility (2/2) — Fully deterministic: all data computed mathematically, no random elements
• CQ-03: Clean Imports (1/2) — library(dplyr) imported but never used anywhere in the code
• CQ-04: Code Elegance (2/2) — Well-organized; Newton-Raphson solver correctly implemented; no fake UI; appropriate complexity
• CQ-05: Output & API (1/1) — Saves plot-{THEME}.png using ragg::agg_png at 8×4.5in, dpi=400 = 3200×1800px

Library Mastery (5/10)

• LM-01: Idiomatic Usage (3/5) — Uses theme_void(), coord_fixed(), geom_polygon(), geom_path() idiomatically; ggplot2 used as a drawing framework
• LM-02: Distinctive Features (2/5) — theme_void() + coord_fixed() are ggplot2-specific; layer composition is characteristic; but projection math is entirely manual with no ggplot2 geo capabilities leveraged

Score Caps Applied

• None

Strengths

• Mathematically rigorous Mollweide projection implemented from scratch (Newton-Raphson solver for theta), entirely within ggplot2 drawing primitives — correctly avoids sf dependency
• Excellent dual-theme adaptation with custom ocean tokens (#C4DCF0 light / #152030 dark) that are semantically appropriate and visually coherent in both renders
• Brand green (#009E73) correctly used for Tissot indicatrices (the sole data series); chrome adapts fully and correctly in both themes
• Clean visual hierarchy: faint graticule (alpha=0.45) → bolder equator/prime meridian (alpha=0.70) → prominent Tissot ellipses — guides the viewer correctly
• Tissot indicatrices correctly demonstrate the equal-area property with proper latitude cosine correction applied to the longitude extent

Weaknesses

• Missing land masses and country borders: spec explicitly requires "world country boundaries" and "clean coastlines and country borders." The plot shows only ocean with no land polygons. Fix: add library(maps) and map_data("world") to get country boundary lon/lat coordinates, project through the existing mollweide() function, render with geom_polygon() (land fill) + geom_path() (borders) before the graticule layer — no sf dependency needed
• library(dplyr) imported but never used — remove it
• Subtitle at 8.5pt may be too small at mobile scale (~400px wide) — consider raising to 9.5–10pt

Issues Found

1. SC-02 PARTIAL: Missing country boundaries and coastlines
   • Fix: library(maps) + map_data("world") → project lon/lat through mollweide() → insert geom_polygon(fill=LAND_BG, color=NA) + geom_path(color=INK_SOFT, linewidth=0.2) between ocean fill and graticule layers
2. CQ-03 MINOR: library(dplyr) unused
   • Fix: Remove from imports
3. DQ-01 PARTIAL: Feature coverage incomplete without land masses — addressed by fix Add workflow diagram for new prototype discovery #1

AI Feedback for Next Attempt

The core projection math is correct and elegant — keep it. The main fix: add library(maps) and use map_data("world") to load country boundary coordinates, project them through the existing mollweide() function, and render as geom_polygon() (land fill, e.g. a warm neutral like "#DDD8CC" for light / "#2A2A25" for dark) + geom_path() (borders, INK_SOFT at linewidth=0.2). Insert these as new layers between the ocean fill and the graticule. Also remove the unused library(dplyr) import. These two changes will substantially improve spec compliance and data quality scores.

Verdict: REJECTED

[github-actions]

🔧 Repair Attempt 1/4

Applied fixes based on AI review feedback.

Status: Repair completed, re-triggering review...

---

🤖 impl-repair

[claude]

AI Review - Attempt 2/3

Image Description

Light render (plot-light.png): The plot displays a Mollweide equal-area world map on a warm off-white #FAF8F1 background. The ocean fill is a muted cornflower blue (#C4DCF0) and land masses are rendered in a sandy beige (#D4C9A8). Country borders and coastlines are drawn as very thin semi-transparent lines in the soft ink color. A 30° graticule grid covers the map, with the equator and prime meridian drawn slightly bolder. Tissot indicatrices (30 circles arranged at 5 latitude bands × 6 longitude positions) are filled in brand green #009E73 — consistent with the anyplot palette position 1. The ellipse boundary of the Mollweide projection is clearly outlined. The title "map-projections · r · ggplot2 · anyplot.ai" is center-aligned in dark ink at 12pt. The subtitle is in secondary ink (#4A4A44) at 9.5pt, though it appears nearly left-aligned in rendering (long text spans ~90% of canvas width). All text is clearly readable against the light background.

Dark render (plot-dark.png): The same map is rendered on a near-black #1A1A17 background. The ocean becomes a deep navy (#152030) and land masses shift to a dark olive green (#2A3020), providing clear land-sea contrast. The Tissot indicatrices remain identical brand green #009E73 — data colors are unchanged between themes as required. The graticule lines and ellipse boundary are rendered in the light INK_SOFT token (#B8B7B0). Title and subtitle use light text tokens (#F0EFE8 / #B8B7B0) and are clearly readable against the dark background. No dark-on-dark text failures detected.

Both paragraphs are required. A review that only describes one render is invalid.

Score: 85/100

Category	Score	Max
Visual Quality	27	30
Design Excellence	12	20
Spec Compliance	14	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	7	10
Total	85	100

Visual Quality (27/30)

• VQ-01: Text Legibility (7/8) — Title and subtitle are readable in both themes. Minor: the subtitle (~90 chars at 9.5pt) spans ~90% of canvas width and appears nearly edge-to-edge, leaving minimal breathing room.
• VQ-02: No Overlap (5/6) — No text or data overlaps. Minor: Tissot circles at high latitudes are visually very close in some longitude bands.
• VQ-03: Element Visibility (6/6) — Tissot indicatrices, land masses, ocean, and graticule are all clearly distinguishable in both themes.
• VQ-04: Color Accessibility (2/2) — Brand green against ocean blue/navy has adequate contrast; CVD-safe palette used.
• VQ-05: Layout & Canvas (3/4) — Canvas is correct 3200×1800. Map ellipse is well-centered. Slight deduction: subtitle text approaches canvas edges without the visual breathing room the 50pt margin was intended to provide.
• VQ-06: Axis Labels & Title (2/2) — theme_void() is correct for a map; subtitle provides descriptive context. No axis labels needed.
• VQ-07: Palette Compliance (2/2) — Tissot indicatrices use ANYPLOT_PALETTE[1] (#009E73). Backgrounds #FAF8F1 (light) / #1A1A17 (dark). Data colors identical across themes.

Design Excellence (12/20)

• DE-01: Aesthetic Sophistication (5/8) — Above default. Theme-adaptive ocean/land colors (not just grey fill), intentional visual hierarchy (equator/prime meridian slightly bolder than graticule), and good use of alpha on borders. Not exceptional but clearly designed.
• DE-02: Visual Refinement (3/6) — theme_void() removes axes appropriately for a map. Subtle graticule opacity and thin border alpha add polish. Ellipse boundary is clean. Missing: the subtitle nearly touching canvas edges is a refinement miss.
• DE-03: Data Storytelling (4/6) — The Tissot indicatrices effectively communicate the equal-area property. The subtitle explicitly explains what the circles prove ("every circle covers identical surface"). Clear focal point and educational narrative.

Spec Compliance (14/15)

• SC-01: Plot Type (5/5) — Correct world map implementation. Spec explicitly allows "a single projection with clear title indicating which one."
• SC-02: Required Features (3/4) — Graticule at 30° intervals ✓, coastlines and country borders ✓, Tissot indicatrices ✓, neutral land color scheme ✓. Minor deduction: spec description emphasizes "demonstrating different cartographic projections" (plural); showing only Mollweide limits the comparative educational value. Single-projection is explicitly allowed in Notes but the description's intent is multiple projections.
• SC-03: Data Mapping (3/3) — Real world boundary data from maps package, correct Mollweide projection math (Newton-Raphson), all country polygons displayed.
• SC-04: Title & Legend (3/3) — Title matches {spec-id} · r · ggplot2 · anyplot.ai exactly. No legend needed (single-series Tissot in brand green, subtitle identifies it).

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — World boundaries, ocean fill, land fill, graticule, equator highlight, prime meridian highlight, Tissot indicatrices all present.
• DQ-02: Realistic Context (5/5) — Uses real world boundary data; Mollweide is a genuine equal-area projection; Tissot indicatrices are a standard cartographic analysis technique. Neutral educational context.
• DQ-03: Appropriate Scale (4/4) — 30° graticule intervals are standard. Tissot radius of 5° at 5 latitude bands × 6 longitude positions is well-calibrated for the canvas.

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Helper functions solve_theta and mollweide are genuinely necessary to implement the projection without sf/ggmap (which are out of scope per library rules). Code structure is flat and readable.
• CQ-02: Reproducibility (2/2) — No random seed needed; all data is deterministic (real map data + mathematical projection).
• CQ-03: Clean Imports (2/2) — ggplot2, maps, ragg — all three used.
• CQ-04: Code Elegance (2/2) — Appropriate ggplot2 layer composition. No fake UI elements.
• CQ-05: Output & API (1/1) — Saves as plot-{THEME}.png using ragg::agg_png. Correct dimensions (8×4.5 in at 400dpi = 3200×1800).

Library Mastery (7/10)

• LM-01: Idiomatic Usage (4/5) — Correct use of theme_void() for a map, coord_fixed() for proper aspect ratio, multi-layer geom composition via ggplot2 grammar. Good idiomatic patterns.
• LM-02: Distinctive Features (3/5) — Implements the Mollweide projection from scratch in pure R math (since sf/ggmap are out of scope), then layers the result through ggplot2's geom system. The layer-composition approach (ocean fill → land fill → borders → graticule → equator/prime meridian → Tissot → boundary outline) showcases ggplot2's grammar of graphics well.

Score Caps Applied

• None

Strengths

• Implements Mollweide projection from scratch using Newton-Raphson iteration — correctly bypasses sf/ggmap (out of scope) without sacrificing accuracy
• Theme-adaptive ocean and land colors (not just grey fill): #C4DCF0/#152030 ocean, #D4C9A8/#2A3020 land — thoughtfully chosen for both surfaces
• Tissot indicatrices correctly demonstrate the equal-area property; subtitle explains the visual insight ("every circle covers identical surface")
• Perfect palette compliance: brand green #009E73 for Tissot circles, correct backgrounds in both themes
• Visual hierarchy: equator and prime meridian are bolder than regular graticule — reads naturally
• Perfect code quality: deterministic, clean imports, idiomatic ggplot2 grammar, correct canvas size

Weaknesses

• Subtitle is ~90 characters long at 9.5pt, causing it to span ~90% of the canvas width and appear nearly edge-to-edge — reduce subtitle length or reduce font size to ~8pt to restore visual breathing room
• Only Mollweide projection is shown; the spec description emphasizes demonstrating multiple projections (Mercator, Robinson, Mollweide, etc.) for comparison — consider a 2×2 faceted layout showing 4 projections side-by-side to fully satisfy the educational intent (each panel labeled with projection name)
• Tissot indicatrices are limited to ±60° latitude; including a row at ±75° or near-polar positions would better illustrate the extreme distortion in those regions

Issues Found

1. SC-02 PARTIAL: Single projection shown vs. spec's emphasis on demonstrating multiple projections for comparison
   • Fix: Add a 2×2 small-multiple layout (e.g., Mercator, Robinson, Mollweide, Equal Earth) with each panel labeled — this requires computing projection math for each type or using facet_wrap with a data frame that includes projection column
2. VQ-05/VQ-01 MINOR: Subtitle too long, approaches canvas edges
   • Fix: Shorten subtitle to e.g. "Mollweide equal-area · Tissot indicatrices confirm equal-area property" (~55 chars) or reduce size to 8pt

AI Feedback for Next Attempt

The implementation is strong technically — the Mollweide math is correct, themes are properly wired, and palette compliance is perfect. To push to 90+: (1) Show multiple projections in a 2×2 grid to match the spec's core intent of "demonstrating different cartographic projections" — compute Mercator, Robinson/Winkel Tripel (approximated), and Mollweide in separate data frames, bind with a projection column, and use facet_wrap; (2) Shorten the subtitle or reduce its size to avoid the near-edge spanning; (3) Consider adding graticule labels (30°N, 60°S, etc.) as annotations to further the educational value.

Verdict: APPROVED