feat(ggplot2): implement acf-pacf

[github-actions]

Implementation: acf-pacf - r/ggplot2

Implements the r/ggplot2 version of acf-pacf.

File: plots/acf-pacf/implementations/r/ggplot2.R

Parent Issue: #4663

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

Light render (plot-light.png): The plot shows two vertically stacked faceted panels on a warm off-white (#FAF8F1) background. The top panel is labelled ACF and the bottom PACF in bold strip headers with a slightly elevated background (#FFFDF6). Stem lines (vertical segments from zero to correlation value) are rendered in brand green (#009E73). Lag 0 ACF reaches 1.0 and is clearly prominent; lags 1 and 2 show moderate positive ACF values (~0.48), while higher lags decay quickly inside the confidence bounds. PACF shows a sharp cutoff after lag 2 (values ~0.48 and ~-0.28) as expected for AR(2). Dashed confidence interval lines appear in a muted gray (INK_MUTED). The shared y-axis is labelled "Autocorrelation" and the shared x-axis is labelled "Lag" with ticks at every 5 lags. Title "acf-pacf · r · ggplot2 · anyplot.ai" is clear at the top. All text — title, axis labels, tick labels, facet strips — is readable in dark ink against the light background. No legibility issues. Legibility verdict: PASS.

Dark render (plot-dark.png): Identical layout on a warm near-black (#1A1A17) background. The title and all axis/tick labels appear in light text (F0EFE8 / B8B7B0) against the dark surface — no dark-on-dark failures observed. Brand green stems (#009E73) are identical in color to the light render, confirming data colors are theme-independent. Strip headers show ACF/PACF in light bold text on a slightly elevated dark background (#242420). Confidence interval dashes appear in a lighter muted tone, still clearly visible. Grid lines are visible but subtle. The plot reads cleanly in dark mode with full legibility. Legibility verdict: PASS.

Both renders pass the theme-readability checklist. No dark-on-dark issues, no edge clipping, no overflow.

Score: 85/100

Category	Score	Max
Visual Quality	28	30
Design Excellence	10	20
Spec Compliance	15	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	7	10
Total	85	100

Visual Quality (28/30)

• VQ-01: Text Legibility (7/8) — Title, axis labels, tick labels, and strip headers all legible in both themes. Slightly conservative from max: stems for near-zero lags are thin (linewidth=0.8) which may strain visibility at mobile scale (~400px).
• VQ-02: No Overlap (6/6) — No text collisions, no data/label overlap.
• VQ-03: Element Visibility (5/6) — Main signal (lags 0–2) is very prominent. Near-zero stems at higher lags are short and thin (linewidth=0.8) but still identifiable at full resolution.
• VQ-04: Color Accessibility (2/2) — Single series in brand green; confidence lines in muted gray. CVD-safe.
• VQ-05: Layout & Canvas (4/4) — 3200×1800 canvas confirmed. Two panels fill the space well, no overflow, no clipping.
• VQ-06: Axis Labels & Title (2/2) — X-axis: "Lag" ✓; Y-axis: "Autocorrelation" (shared, appropriate for faceted layout); facet strips serve as panel-specific labels; title correct.
• VQ-07: Palette Compliance (2/2) — Data: #009E73 (Imprint position 1). Backgrounds: #FAF8F1 light / #1A1A17 dark. Chrome tokens correct in both renders.

Design Excellence (10/20)

• DE-01: Aesthetic Sophistication (4/8) — Correct Imprint palette, proper theme tokens, clean layout. Professional but not exceptional. No additional decorative design choices beyond correct defaults.
• DE-02: Visual Refinement (3/6) — Minor grid blank, subtle major grid (linewidth=0.2), strip background uses elevated token, panel border is clean. Slightly above default; spines retained on all four sides (acceptable for faceted panels).
• DE-03: Data Storytelling (3/6) — Excellent data choice: AR(2) with phi=(0.7, −0.3) produces the canonical textbook ACF geometric-decay + PACF sharp-cutoff pattern, making the plot genuinely instructive for ARIMA model-order identification.

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — Correct ACF/PACF stem plot using geom_segment and facet_wrap.
• SC-02: Required Features (4/4) — Two stacked panels ✓, vertical stem lines from zero ✓, 95% CI dashed lines ✓, lag 0 in ACF ✓, PACF starts at lag 1 ✓.
• SC-03: Data Mapping (3/3) — X: lag numbers 0–36; Y: correlation values with free y-scales per panel; ACF top, PACF bottom.
• SC-04: Title & Legend (3/3) — Title "acf-pacf · r · ggplot2 · anyplot.ai" ✓. No legend needed (single series).

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — Both ACF and PACF shown; lag 0 behavior; significance indicators; classic AR(2) pattern with clear structure.
• DQ-02: Realistic Context (5/5) — AR(2) with realistic parameters (phi1=0.7, phi2=−0.3), 300 observations — appropriate for reliable correlation estimates. Domain-neutral.
• DQ-03: Appropriate Scale (4/4) — 36 lags (within 30–40 target); y-axis scales to data range per panel.

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Flat script, no functions or classes.
• CQ-02: Reproducibility (2/2) — set.seed(42); arima.sim is deterministic given seed.
• CQ-03: Clean Imports (2/2) — Only ggplot2 and ragg imported; both used.
• CQ-04: Code Elegance (2/2) — Clean data construction via rbind of ACF/PACF data frames; factor level ordering ensures correct facet order; no fake UI.
• CQ-05: Output & API (1/1) — sprintf("plot-%s.png", THEME) ✓; ragg::agg_png ✓.

Library Mastery (7/10)

• LM-01: Idiomatic Usage (4/5) — facet_wrap(~ type, ncol=1, scales="free_y") for two-panel layout ✓; geom_segment for stems ✓; geom_hline for reference lines ✓; tidy long-form data for faceting ✓; scale_x_continuous(breaks=...) for tick control ✓.
• LM-02: Distinctive Features (3/5) — Uses ggplot2-distinctive faceting with per-strip styled backgrounds (strip.background + elevated token), free y-scales, and the tidy data grammar pattern (rbind ACF/PACF into a single long data frame with a type factor).

Score Caps Applied

• None

Strengths

• Excellent data choice: AR(2) with phi=(0.7, −0.3) produces the canonical textbook ACF/PACF pattern — geometric decay in ACF, sharp cutoff at lag 2 in PACF — making the plot genuinely instructive
• Full spec compliance: all required features (stems, CI bands, lag 0 in ACF, PACF from lag 1) correctly implemented
• Theme adaptation is flawless: all chrome tokens correctly flip; data colors identical across renders
• Clean tidy-data approach: combined data frame with type factor drives faceting elegantly
• Perfect code quality: flat script, seeded, no unused imports

Weaknesses

• Stem linewidth (0.8) is slightly thin for near-zero correlations at mobile scale; increasing to 1.0–1.2 would improve visibility of small-but-significant stems
• Design Excellence is functional but minimal: no annotation to highlight the significant lags (e.g., a subtle label on lags 1–2 indicating "order p=2"), no color differentiation for stems above/below the confidence band
• All four panel borders are shown (panel.border keeps all sides); removing top and right borders per style-guide default would improve minimalism

Issues Found

1. DE-01 LOW: No design choices beyond correct defaults — no annotations, no emphasis on significant lags, no visual hierarchy within the data
   • Fix: Add a distinct fill or size emphasis to stems that exceed the CI bounds, or annotate lag 2 as the AR order cutoff
2. VQ-03 MINOR: Stem linewidth=0.8 produces very thin lines for near-zero correlations
   • Fix: Increase to linewidth=1.2 for better visibility at thumbnail/mobile sizes

AI Feedback for Next Attempt

Increase stem linewidth to 1.2 for better visibility. Consider adding visual emphasis to statistically significant stems (those exceeding the CI bands) by coloring them differently (e.g., full Imprint green for significant, muted gray for insignificant). Remove top and right panel borders to align with the style-guide L-spine default. These changes would push Design Excellence from 10/20 toward 14–15/20.

Verdict: APPROVED

[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): Warm off-white #FAF8F1 background. Title "acf-pacf · r · ggplot2 · anyplot.ai" in dark ink at top-left — clearly readable. Two facet panels ("ACF" top, "PACF" bottom) with bold strip labels on elevated cream #FFFDF6 backgrounds with subtle INK_SOFT borders. Y-axis label "Autocorrelation" and X-axis label "Lag" in dark ink, both readable. Tick labels in INK_SOFT at 8pt — readable. Significant lags colored in brand green #009E73 (lag 0 in ACF at 1.0, lag 1 in both panels, lag 2 in PACF), non-significant lags in muted gray (INK_MUTED). Dashed CI lines at ±0.113 in muted gray. Subtle major grid, no minor grid, L-shaped spines. All text clearly readable against light background — no light-on-light failures.

Dark render (plot-dark.png): Warm near-black #1A1A17 background. Title and all axis/tick labels rendered in light ink (#F0EFE8 / #B8B7B0) — clearly readable against the dark surface. Facet strips use ELEVATED_BG #242420 with light-colored strip text — correct elevated treatment. Grid lines appear as subtle light gray lines against the dark background. Data colors are identical to the light render: significant lags still #009E73, non-significant still medium gray — only chrome flipped. No dark-on-dark failures detected. Brand green fully visible. Both renders theme-correct.

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

Score: 88/100

Category	Score	Max
Visual Quality	27	30
Design Excellence	13	20
Spec Compliance	15	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	8	10
Total	88	100

Visual Quality (27/30)

• VQ-01: Text Legibility (7/8) — All font sizes explicitly set; readable in both themes. Minor: stem linewidth=1.2 slightly thin for 3200×1800 canvas but text itself is fine.
• VQ-02: No Overlap (6/6) — No overlapping text or data elements in either render
• VQ-03: Element Visibility (4/6) — Significant stems clearly visible in green; non-significant stems at linewidth=1.2 are thin on a 3200×1800 canvas — visible but not optimally sized
• VQ-04: Color Accessibility (2/2) — Green #009E73 vs gray is CVD-safe and distinguishable without relying on hue alone
• VQ-05: Layout & Canvas (4/4) — Two-panel facet fills canvas well, balanced margins, no wasted space
• VQ-06: Axis Labels & Title (2/2) — "Autocorrelation" / "Lag"; facet strips label panels; title format correct
• VQ-07: Palette Compliance (2/2) — Significant series uses #009E73; backgrounds #FAF8F1 / #1A1A17; data colors identical across themes; chrome theme-adaptive

Design Excellence (13/20)

• DE-01: Aesthetic Sophistication (5/8) — Intentional semantic color (significant=green, non-significant=gray), clean elevated strip backgrounds — clearly above defaults, not quite publication-ready
• DE-02: Visual Refinement (4/6) — L-shaped spines, minor grid removed, subtle major grid; strip INK_SOFT border slightly heavy for maximum refinement
• DE-03: Data Storytelling (4/6) — Color-coded significance creates clear visual hierarchy; AR(2) ACF decay + PACF cutoff at lag 2 tells a textbook-quality story

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — ACF top + PACF bottom via facet_wrap; vertical stems from zero; shared lag axis; dashed CI lines
• SC-02: Required Features (4/4) — Lag 0 in ACF; PACF from lag 1; CI at ±1.96/√300; 36 lags (within 30–40 range)
• SC-03: Data Mapping (3/3) — X=lag, Y=correlation; both panels correctly mapped
• SC-04: Title & Legend (3/3) — Title "acf-pacf · r · ggplot2 · anyplot.ai" matches required format; guide=\"none\" appropriate for two-class semantic coloring

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — Shows ACF geometric decay and PACF cutoff; significant/non-significant lags; 300 observations
• DQ-02: Realistic Context (5/5) — AR(2) synthetic series — standard, neutral, real-world-plausible time series example
• DQ-03: Appropriate Scale (4/4) — AR(2) with φ₁=0.7, φ₂=−0.3; CI at ±1.96/√300 ≈ ±0.113 — mathematically correct

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Clean linear flow: imports → tokens → data → plot → save
• CQ-02: Reproducibility (2/2) — set.seed(42) present
• CQ-03: Clean Imports (2/2) — Only ggplot2 and ragg imported, both used
• CQ-04: Code Elegance (2/2) — Idiomatic geom_segment + aes(color=significant); no fake functionality
• CQ-05: Output & API (1/1) — Saves as plot-%s.png; uses linewidth= (not deprecated size=) for ggplot2 ≥3.4

Library Mastery (8/10)

• LM-01: Idiomatic Usage (5/5) — Grammar-of-graphics aesthetic mapping of computed boolean to color; facet_wrap with scales=\"free_y\"; proper theme_minimal base + theme layering
• LM-02: Distinctive Features (3/5) — Uses ggplot2-distinctive aes(color=significant) data-driven mapping and faceting with scale independence — recognizable grammar approach; could go further (e.g., annotate for CI band, coord_flip for alternate layout)

Score Caps Applied

• None

Strengths

• Significant-lag color coding (green #009E73 vs gray INK_MUTED) elegantly tells the ACF/PACF story at a glance
• Correct facet_wrap design with free_y scales and clearly labeled strips
• Full spec compliance: lag 0 in ACF, PACF from lag 1, correct CI formula ±1.96/√N, 36 lags
• AR(2) synthetic data with φ₁=0.7, φ₂=−0.3 produces textbook-quality ACF/PACF patterns
• Proper theme-adaptive chrome: L-shaped spines, subtle grid, ELEVATED_BG strip backgrounds
• Perfect code structure: KISS, set.seed(42), clean imports, correct linewidth API

Weaknesses

• Stem linewidth=1.2 is thin for a 3200×1800 canvas — non-significant bars are very short and barely readable; increase to linewidth=1.6–2.0 for better visibility
• Strip border (color=INK_SOFT on strip.background) creates a slightly heavy box frame; consider element_rect(fill=ELEVATED_BG, color=NA) to remove the border for a cleaner elevated look
• LM-02: the implementation does not exploit ggplot2 grammar beyond faceting + geom_segment; consider using ggplot2-idiomatic annotate or geom_ribbon for the CI band

Issues Found

1. VQ-03 MINOR: Stem lines at linewidth=1.2 render thin at 3200×1800; non-significant (short, small-value) bars are barely visible
   • Fix: Increase geom_segment linewidth to 1.6–2.0
2. DE-02 MINOR: Strip background border slightly heavy, reducing visual cleanliness
   • Fix: Change strip.background = element_rect(fill=ELEVATED_BG, color=NA) to remove the border

AI Feedback for Next Attempt

The implementation is strong — full spec compliance, perfect code quality, and effective data storytelling through significance coloring. To push higher: (1) increase stem linewidth to ~1.8 so non-significant bars read clearly at canvas resolution; (2) remove the strip.background border for a cleaner facet look (color=NA); (3) for LM-02, consider using geom_ribbon or a filled confidence band instead of two geom_hline lines to leverage more of ggplot2's grammar.

Verdict: APPROVED