feat(matplotlib): implement residual-plot

[github-actions]

Implementation: residual-plot - matplotlib

Implements the matplotlib version of residual-plot.

File: plots/residual-plot/implementations/matplotlib.py

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

The plot displays a residual plot with fitted values (ranging from approximately -5 to 50) on the x-axis and residuals (observed minus predicted, ranging from about -10 to 10) on the y-axis. The title reads "residual-plot · matplotlib · pyplots.ai" in large font at the top. Regular residual points are shown as blue circles (#306998) with white edges, while outliers exceeding 2 standard deviations are highlighted in yellow (#FFD43B) with blue edges. A solid black horizontal reference line at y=0 represents perfect predictions. Dashed gray horizontal lines mark the ±2σ boundaries, with a light blue shaded band between them. A red polynomial trend line curves through the data, showing a clear U-shaped pattern indicating the non-linearity that the linear model fails to capture. The legend is positioned in the upper left corner with clear labels for all elements.

Quality Score: 93/100

Criteria Checklist

Visual Quality (37/40 pts)

• VQ-01: Text Legibility (10/10) - Title at 24pt, axis labels at 20pt, tick labels at 16pt - all perfectly readable
• VQ-02: No Overlap (8/8) - No overlapping text anywhere
• VQ-03: Element Visibility (7/8) - Markers sized appropriately (s=150) for 150 data points with good alpha (0.7). Slightly larger than optimal per guidelines (100-200 for 100-300 points) but still effective
• VQ-04: Color Accessibility (5/5) - Blue and yellow color scheme is colorblind-safe, good contrast between regular points and outliers
• VQ-05: Layout Balance (5/5) - Plot fills canvas well with balanced margins
• VQ-06: Axis Labels (1/2) - "Fitted Values" and "Residuals (Observed - Predicted)" are descriptive but lack units
• VQ-07: Grid & Legend (2/2) - Grid is subtle (alpha=0.3), legend well-placed in upper left with good framealpha

Spec Compliance (25/25 pts)

• SC-01: Plot Type (8/8) - Correct residual plot with residuals vs fitted values
• SC-02: Data Mapping (5/5) - X-axis shows fitted values, Y-axis shows residuals (y_true - y_pred)
• SC-03: Required Features (5/5) - All spec features present: horizontal reference line at y=0, alpha transparency, outlier coloring (>2σ), ±2σ bands, trend/smoothing line to detect patterns
• SC-04: Data Range (3/3) - All data visible within axis limits
• SC-05: Legend Accuracy (2/2) - Legend labels correctly describe all plot elements
• SC-06: Title Format (2/2) - Correctly uses "residual-plot · matplotlib · pyplots.ai"

Data Quality (18/20 pts)

• DQ-01: Feature Coverage (7/8) - Excellent demonstration: shows non-linear pattern in residuals (U-shape from quadratic component in true data), outliers visible (4-5 yellow points), heteroscedasticity somewhat visible. Could show more varied outlier distribution
• DQ-02: Realistic Context (7/7) - Realistic regression scenario with simulated linear model fit to data with quadratic component
• DQ-03: Appropriate Scale (4/5) - Values are reasonable; fitted values 0-50 and residuals ±10 are plausible for regression diagnostics

Code Quality (10/10 pts)

• CQ-01: KISS Structure (3/3) - Clean script: imports → data → plot → save, no functions/classes
• CQ-02: Reproducibility (3/3) - Uses np.random.seed(42)
• CQ-03: Clean Imports (2/2) - Only matplotlib.pyplot and numpy imported, both used
• CQ-04: No Deprecated API (1/1) - No deprecated functions used
• CQ-05: Output Correct (1/1) - Saves as 'plot.png'

Library Features (3/5 pts)

• LF-01: Uses distinctive library features (3/5) - Good use of matplotlib's axhline, fill_between, scatter with styling, and polyfit for trend line. Could leverage more advanced features like subplot with marginal distributions or custom tick formatting

Strengths

• Excellent implementation of all spec requirements including reference line, ±2σ bands, outlier highlighting, and trend line
• Deliberately generated data with quadratic component to demonstrate non-linear residual pattern detection
• Clean, well-structured code following KISS principles with clear data generation logic
• Good visual design with colorblind-safe colors (blue/yellow) and appropriate transparency levels
• Comprehensive legend explaining all visual elements

Weaknesses

• Axis labels could include units or more context (e.g., "Fitted Values (ŷ)" or similar notation)
• Could demonstrate heteroscedasticity more clearly with variance that changes across fitted values

Verdict: APPROVED