feat(matplotlib): implement scatter-lag

[github-actions]

Implementation: scatter-lag - matplotlib

Implements the matplotlib version of scatter-lag.

File: plots/scatter-lag/implementations/matplotlib.py

Parent Issue: #5251

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

The plot shows a lag-1 scatter plot of a synthetic AR(1) time series (500 observations, φ=0.85). The x-axis shows y(t) ranging from approximately -4.5 to 5.5, and the y-axis shows y(t+1) over the same range. Points are colored by time index using the viridis colormap (purple=early, yellow=late), with a vertical colorbar on the right labeled "Time Index" (range 0–500). A dashed gray diagonal reference line (y=x) runs from bottom-left to top-right. In the upper-left corner, an annotation reads "r = 0.834". The scatter cloud forms a clear elongated ellipse along the diagonal, demonstrating strong positive autocorrelation. The top and right spines are removed; a subtle full-grid is visible. The title reads "AR(1) Autocorrelation · scatter-lag · matplotlib · pyplots.ai". Overall layout is clean and balanced on a 16:9 canvas.

Score: 84/100

Category	Score	Max
Visual Quality	26	30
Design Excellence	12	20
Spec Compliance	15	15
Data Quality	13	15
Code Quality	10	10
Library Mastery	8	10
Total	84	100

Visual Quality (26/30)

• VQ-01: Text Legibility (7/8) — All key fonts explicitly set (title 24pt, labels 20pt, ticks 16pt, r-annotation 20pt); colorbar tick labels at 14pt are slightly below the 16pt minimum
• VQ-02: No Overlap (6/6) — No text collisions; r annotation, colorbar, and axis labels are well separated
• VQ-03: Element Visibility (3/6) — 500 points with s=120, alpha=0.65 produces significant overplotting in the dense center; guidelines recommend s=20–50, alpha=0.3–0.5 for 300+ points
• VQ-04: Color Accessibility (4/4) — Viridis is perceptually uniform and colorblind-safe
• VQ-05: Layout & Canvas (4/4) — 16:9 canvas, balanced margins, colorbar well-placed
• VQ-06: Axis Labels & Title (2/2) — "y(t)" and "y(t + 1)" are standard mathematical notation for a lag plot; no applicable units for synthetic dimensionless data

Design Excellence (12/20)

• DE-01: Aesthetic Sophistication (4/8) — Well-configured library defaults: viridis is matplotlib's recommended sequential colormap, spine removal follows the guide. Does not yet reach "strong design" territory
• DE-02: Visual Refinement (4/6) — Spines removed, grid alpha=0.2 is subtle; minor issue: full x+y grid rather than y-only as recommended in the library guide
• DE-03: Data Storytelling (4/6) — Temporal color encoding reveals how autocorrelation evolves over time; diagonal reference line + r annotation together create a clear narrative; visual hierarchy is present though not exceptional

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — Correct lag scatter plot (y(t) vs y(t+k))
• SC-02: Required Features (4/4) — Diagonal reference line, time-index color coding, correlation annotation, and configurable lag variable all present
• SC-03: Data Mapping (3/3) — x=y(t), y=y(t+1) correctly assigned; axes cover full data range
• SC-04: Title & Legend (3/3) — Title contains spec-id · library · pyplots.ai; colorbar labeled "Time Index" is appropriate; no discrete legend needed for continuous color scale

Data Quality (13/15)

• DQ-01: Feature Coverage (5/6) — Demonstrates strong positive autocorrelation clearly; could improve by showing contrast (e.g., a second lag or annotation explaining what the diagonal pattern means)
• DQ-02: Realistic Context (4/5) — Synthetic AR(1) process is explicitly listed in the spec as a valid example; neutral and appropriate, though abstract without real-world labeling
• DQ-03: Appropriate Scale (4/4) — AR(1) with φ=0.85 and unit noise produces values in [-5, 5] — correct and realistic

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Linear: imports → data generation → correlation → plot → save; no functions or classes
• CQ-02: Reproducibility (2/2) — np.random.seed(42) set
• CQ-03: Clean Imports (2/2) — Only matplotlib.pyplot and numpy imported, both used
• CQ-04: Code Elegance (2/2) — Clean, idiomatic Python; appropriate complexity
• CQ-05: Output & API (1/1) — Saves as plot.png with dpi=300; uses ax methods throughout

Library Mastery (8/10)

• LM-01: Idiomatic Usage (5/5) — ax.scatter with c= + cmap, fig.colorbar(scatter, ax=ax), ax.transAxes for annotation, tight_layout — all idiomatic matplotlib patterns
• LM-02: Distinctive Features (3/5) — Continuous color mapping via scatter c= + colorbar is a recognizable matplotlib pattern; white edgecolors on markers is a nice touch, but overall usage is not uniquely distinctive

Score Caps Applied

• None — no caps triggered (DE-01=4, DE-02=4 → cap requires both ≤ 2)

Strengths

• Perfect spec compliance: all required features implemented (diagonal line, temporal coloring, r annotation, configurable lag)
• Excellent code quality: clean KISS structure, reproducible, idiomatic matplotlib usage
• Good data storytelling: the viridis time-index coloring meaningfully encodes temporal evolution

Weaknesses

• Marker size too large for 500 points (s=120 vs recommended s=20–50); overplotting obscures density in the central cluster
• Design is well-configured but not elevated: viridis + spine removal follows the guide but doesn't add creative aesthetic lift
• Full x+y grid should be y-only per library conventions; colorbar ticks at 14pt should be ≥16pt

Issues Found

1. VQ-03 LOW: s=120 is 3× larger than recommended for 500 data points, causing significant overplotting in the dense diagonal band
   • Fix: Reduce to s=25–40 and alpha=0.35–0.45
2. DE-01 LOW: Design follows the library guide but lacks creative differentiation
   • Fix: Consider a custom sequential palette (e.g., plasma or a custom dark-to-bright scheme), slightly larger reference line, or a background hue to frame the plot area
3. DE-02 MINOR: Full grid adds visual noise for a scatter plot
   • Fix: Replace ax.grid(True, ...) with ax.yaxis.grid(True, alpha=0.2, linewidth=0.8)
4. VQ-01 MINOR: Colorbar tick labels at 14pt are below the 16pt minimum
   • Fix: Change cbar.ax.tick_params(labelsize=14) to labelsize=16

AI Feedback for Next Attempt

Fix marker sizing first (s=25–40, alpha=0.35–0.45 for 500 points) to reduce overplotting. Switch to y-only grid. Raise colorbar tick labels to 16pt. For design lift, try a warmer sequential colormap (plasma or magma) and consider adding a subtle fill or background to the scatter region to emphasize the diagonal correlation band. These changes should push Visual Quality and Design Excellence above current levels.

Verdict: REJECTED

[github-actions]

🔧 Repair Attempt 1/3

Applied fixes based on AI review feedback.

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

---

🤖 impl-repair

[claude]

AI Review - Attempt 2/3

Image Description

The plot shows a scatter plot of an AR(1) time series against its lag-1 version. The x-axis is labeled "y(t)" and the y-axis "y(t + 1)", with values ranging roughly from −4.5 to 5.5 on both axes. Approximately 499 data points are colored using the viridis colormap, transitioning from dark purple (time index 0) to bright yellow (time index ~499), with a colorbar on the right labeled "Time Index". A dashed diagonal reference line (y = x) runs from lower-left to upper-right in a light gray. The data forms a strong linear cluster along the diagonal, visually confirming high positive autocorrelation. An annotation in the upper-left corner shows "r = 0.834" in bold. The title reads "AR(1) Autocorrelation · scatter-lag · matplotlib · pyplots.ai". Top and right spines are removed; a subtle grid (alpha=0.2) is visible. The overall layout is clean and professional on a light background.

Score: 85/100

Category	Score	Max
Visual Quality	28	30
Design Excellence	12	20
Spec Compliance	14	15
Data Quality	14	15
Code Quality	10	10
Library Mastery	7	10
Total	85	100

Visual Quality (28/30)

• VQ-01: Text Legibility (8/8) — All sizes explicitly set: title=24, labels=20, ticks=16, colorbar=18/14, annotation=20
• VQ-02: No Overlap (6/6) — No overlapping text or elements
• VQ-03: Element Visibility (4/6) — s=120, alpha=0.65 for 499 points; markers are too large for this density (guideline: s=20–50, alpha=0.3–0.5 for 300+), causing visible crowding in the dense center
• VQ-04: Color Accessibility (4/4) — Viridis colormap is perceptually uniform and colorblind-safe
• VQ-05: Layout & Canvas (4/4) — 16:9 aspect ratio, good canvas utilization, colorbar well-positioned
• VQ-06: Axis Labels & Title (2/2) — y(t) and y(t + 1) are the standard notation for lag plots; appropriate

Design Excellence (12/20)

• DE-01: Aesthetic Sophistication (4/8) — Clean and functional but relies on viridis defaults; the r annotation box and spine removal add some polish but the overall design is a well-configured library default
• DE-02: Visual Refinement (4/6) — Top/right spines removed, subtle grid (alpha=0.2), tight_layout; deducted for using full x+y grid instead of the preferred y-axis-only grid
• DE-03: Data Storytelling (4/6) — Temporal coloring reveals time structure within the scatter, r annotation quantifies the visual pattern, diagonal reference line guides interpretation — good visual narrative for a lag plot

Spec Compliance (14/15)

• SC-01: Plot Type (5/5) — Correct lag plot: scatter of y(t) vs y(t+k)
• SC-02: Required Features (4/4) — Diagonal y=x reference line, time-index color encoding, r-value annotation, configurable lag variable all present
• SC-03: Data Mapping (3/3) — x=y(t), y=y(t+1), correct per spec
• SC-04: Title & Legend (2/3) — Title contains all required elements but format is "AR(1) Autocorrelation · scatter-lag · matplotlib · pyplots.ai" instead of the expected {spec-id} · {library} · pyplots.ai starting with the spec-id

Data Quality (14/15)

• DQ-01: Feature Coverage (5/6) — Strong autocorrelation case clearly demonstrated; missing contrast with the random/uncorrelated case which the spec explicitly describes
• DQ-02: Realistic Context (5/5) — AR(1) synthetic process is explicitly listed in the spec as an example; neutral and scientifically appropriate
• DQ-03: Appropriate Scale (4/4) — Values ≈ −4.5 to 5.5 are realistic for a zero-mean AR(1) with Gaussian noise std=1

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Imports → Data → Plot → Save, no functions or classes
• CQ-02: Reproducibility (2/2) — np.random.seed(42) set
• CQ-03: Clean Imports (2/2) — Only matplotlib.pyplot and numpy, both used
• CQ-04: Code Elegance (2/2) — Clean, Pythonic, appropriate complexity for the AR(1) generation
• CQ-05: Output & API (1/1) — Saves as plot.png, dpi=300, bbox_inches='tight', no deprecated functions

Library Mastery (7/10)

• LM-01: Idiomatic Usage (4/5) — Uses ax.scatter with c= colormap, fig.colorbar(), ax methods throughout; correct but doesn't leverage any advanced matplotlib patterns
• LM-02: Distinctive Features (3/5) — The c= parameter with colormap + colorbar on scatter is a distinctive matplotlib pattern; ax.transAxes for relative text placement is a nice touch; still fairly reproducible in seaborn/plotly

Score Caps Applied

• None — no cap conditions triggered

Strengths

• Perfect code quality: clean KISS structure, seed set, correct API usage throughout
• Excellent spec compliance: all required features (diagonal, color-by-time, r annotation) implemented and working
• Temporal color encoding via viridis clearly reveals time structure in the scatter
• r=0.834 annotation with ax.transAxes relative positioning is well-executed
• Font sizes explicitly set at all levels for high-res output

Weaknesses

• Marker size s=120 with alpha=0.65 is too large for 499 data points, causing dense overlap in the center cluster
• Title format deviates from {spec-id} · {library} · pyplots.ai — has extra "AR(1) Autocorrelation ·" prefix
• Full x+y gridlines instead of preferred y-axis-only grid
• Design excellence ceiling limited: no custom background treatment, annotation box style is generic, no additional typographic refinement

Issues Found

1. VQ-03 LOW: Marker size s=120 too large for 499 points
   • Fix: Reduce to s=50–60, alpha=0.45–0.55 to reduce density overlap while preserving the viridis color mapping
2. SC-04: Title format should start with the spec-id
   • Fix: Change to scatter-lag · matplotlib · pyplots.ai (or add a descriptive subtitle after: Lag-1 Autocorrelation (AR(1)) · scatter-lag · matplotlib · pyplots.ai)
3. DE-02: Full grid instead of y-axis only
   • Fix: Replace ax.grid(True, alpha=0.2, linewidth=0.8) with ax.yaxis.grid(True, alpha=0.2, linewidth=0.8)
4. DE-01: Generic annotation box style
   • Fix: Style the r annotation with a light background patch (bbox=dict(boxstyle='round,pad=0.3', facecolor='white', alpha=0.7, edgecolor='none')) and consider a slightly refined color for the diagonal line

AI Feedback for Next Attempt

Fix title format to start with spec-id. Reduce marker size to s=50–60 with alpha=0.45 for better density handling with 499 points. Switch to y-axis-only grid. Refine the r-annotation box with a subtle white background and no border edge. These targeted fixes address the remaining gaps and should push the score above 90.

Verdict: REJECTED