feat(altair): implement andrews-curves

[github-actions]

Implementation: andrews-curves - altair

Implements the altair version of andrews-curves.

File: plots/andrews-curves/implementations/altair.py

Parent Issue: #2859

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

The plot displays Andrews curves for the Iris dataset with three species (Setosa, Versicolor, Virginica) shown in distinct colors. Setosa curves are shown in olive/yellow-green (#6B8E23), Versicolor in yellow (#FFD43B), and Virginica in blue (#306998). The x-axis shows "t (radians)" ranging from approximately -3.4 to 3.4 (roughly -π to π). The y-axis shows "Andrews Curve Value" ranging from about -5 to 6. Each species contains approximately 50 curves (one per observation), with transparency applied to reveal density patterns. The three species show distinct wave patterns: Setosa curves tend to stay in the middle range with a characteristic upward trend at the right, Versicolor shows more spread, and Virginica extends to higher values on the right side. The title reads "Iris Classification · andrews-curves · altair · pyplots.ai" at the top. A legend on the right identifies the three species.

Quality Score: 87/100

Criteria Checklist

Visual Quality (35/40 pts)

• VQ-01: Text Legibility (9/10) - Title and axis labels are readable, though legend text could be slightly larger
• VQ-02: No Overlap (8/8) - No overlapping text elements
• VQ-03: Element Visibility (7/8) - Curves are visible with good transparency, though some areas are dense
• VQ-04: Color Accessibility (4/5) - Blue/yellow/green palette is mostly colorblind-safe, though blue and green could be confused by some
• VQ-05: Layout Balance (5/5) - Good proportions, plot fills canvas well
• VQ-06: Axis Labels (1/2) - Descriptive labels but no units for y-axis (radians shown for x)
• VQ-07: Grid & Legend (1/2) - No visible grid (could help readability), legend is present and positioned well

Spec Compliance (24/25 pts)

• SC-01: Plot Type (8/8) - Correct Andrews curves visualization with Fourier series transformation
• SC-02: Data Mapping (5/5) - Multivariate data correctly mapped to curves via Fourier coefficients
• SC-03: Required Features (5/5) - All spec features present: normalized data, color by category, transparency, t range -π to π
• SC-04: Data Range (3/3) - Axes show full data range
• SC-05: Legend Accuracy (2/2) - Legend labels correctly identify the three species
• SC-06: Title Format (1/2) - Close to required format but uses "Iris Classification" prefix instead of just spec-id

Data Quality (18/20 pts)

• DQ-01: Feature Coverage (7/8) - Shows all three species with clear separation patterns, could show outliers more distinctly
• DQ-02: Realistic Context (7/7) - Iris dataset is a classic, well-known scientific dataset
• DQ-03: Appropriate Scale (4/5) - Normalized values are appropriate, though range -5 to 6 is reasonable for standardized data

Code Quality (7/10 pts)

• CQ-01: KISS Structure (3/3) - Clean linear structure: imports → data → transform → plot → save
• CQ-02: Reproducibility (1/3) - No explicit random seed (iris data is deterministic, but best practice is to include seed)
• CQ-03: Clean Imports (2/2) - All imports are used
• CQ-04: No Deprecated API (1/1) - Uses current Altair API
• CQ-05: Output Correct (0/1) - Saves as plot.png (correct) but also saves plot.html (fine, not penalized)

Library Features (3/5 pts)

• LF-01: Uses distinctive library features (3/5) - Uses Altair's declarative encoding well (mark_line, encode with detail for grouping), but doesn't use interactivity features that Altair is known for

Strengths

• Excellent Andrews curves transformation with correct Fourier series implementation
• Good use of transparency (opacity=0.4) for overlapping curves as recommended in spec
• Proper data normalization using StandardScaler before transformation
• Clear cluster separation visible between species groups
• Correct title format following pyplots.ai convention

Weaknesses

• Missing fixed random seed for full reproducibility (sklearn data is deterministic but explicit seed is preferred)
• Legend text is slightly small relative to plot size
• Could benefit from a subtitle explaining the visualization technique

Verdict: REJECTED

Score of 87 is below the 90 threshold for automatic approval. The implementation is solid but needs minor improvements to reach publication quality.

[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 displays Andrews curves for the Iris dataset with 150 observations transformed into Fourier series curves. Three iris species are color-coded: Setosa in yellow/gold (#FFD43B), Versicolor in blue (#306998), and Virginica in olive green (#6B8E23). The x-axis shows "t (radians)" ranging from approximately -π to π (-3.4 to 3.2), and the y-axis displays "Andrews Curve Value" ranging from about -4 to 6. The title follows the required format: "Iris Classification · andrews-curves · altair · pyplots.ai" with a descriptive subtitle. The legend is positioned in the upper right corner. The curves demonstrate clear visual separation between species, with Setosa (yellow) forming a distinct cluster near y=0, while Versicolor and Virginica show overlapping wave patterns at higher amplitudes.

Quality Score: 91/100

Criteria Checklist

Visual Quality (36/40 pts)

• VQ-01: Text Legibility (9/10) - Title, axis labels, and legend are clearly readable. Font sizes are appropriate for the canvas size. Minor: tick labels could be slightly larger.
• VQ-02: No Overlap (8/8) - No overlapping text elements. All labels and legend are well-positioned.
• VQ-03: Element Visibility (7/8) - Curves are visible with appropriate opacity (0.4). strokeWidth=2 works well for the data density. Very slight difficulty distinguishing individual curves in dense regions.
• VQ-04: Color Accessibility (4/5) - Blue, yellow, and olive green provide reasonable distinction. Yellow on white background could be slightly challenging in some contexts.
• VQ-05: Layout Balance (5/5) - Plot fills canvas well, balanced margins, legend is appropriately positioned near the data.
• VQ-06: Axis Labels (2/2) - "t (radians)" includes units, "Andrews Curve Value" is descriptive.
• VQ-07: Grid & Legend (1/2) - No visible grid (which is acceptable for this plot type), legend is well-placed but could benefit from a background for better contrast.

Spec Compliance (25/25 pts)

• SC-01: Plot Type (8/8) - Correct Andrews curves implementation using Fourier series transformation.
• SC-02: Data Mapping (5/5) - X=t parameter, Y=Andrews curve value, correctly implemented.
• SC-03: Required Features (5/5) - All spec features present: normalized variables, transparency (alpha=0.4), color by category, t range from -π to π.
• SC-04: Data Range (3/3) - Axes show all data appropriately.
• SC-05: Legend Accuracy (2/2) - Legend correctly labels all three species.
• SC-06: Title Format (2/2) - Title includes spec-id "andrews-curves", library "altair", and "pyplots.ai".

Data Quality (18/20 pts)

• DQ-01: Feature Coverage (7/8) - Shows cluster separation between species, demonstrates the Fourier transformation well. Could potentially show outlier detection more explicitly.
• DQ-02: Realistic Context (7/7) - Iris dataset is a classic, appropriate example for Andrews curves as mentioned in the spec.
• DQ-03: Appropriate Scale (4/5) - Standardized data produces sensible curve ranges. 150 observations is within the recommended 30-150 range.

Code Quality (9/10 pts)

• CQ-01: KISS Structure (3/3) - Clean structure: imports → data → transformation → plot → save.
• CQ-02: Reproducibility (3/3) - np.random.seed(42) is set.
• CQ-03: Clean Imports (2/2) - All imports are used (altair, numpy, pandas, sklearn).
• CQ-04: No Deprecated API (1/1) - Uses current Altair API.
• CQ-05: Output Correct (0/1) - Saves as plot.png and plot.html (correct for Altair).

Library Features (3/5 pts)

• LF-01: Uses distinctive library features (3/5) - Uses Altair's declarative encoding with detail for grouping curves, proper mark_line with opacity. Could leverage more Altair-specific features like tooltips or interactive selection.

Strengths

• Correct Fourier series implementation for Andrews curves transformation
• Clear visual separation between iris species demonstrating the clustering capability
• Proper data normalization using StandardScaler as recommended in spec
• Clean, declarative Altair code with appropriate encoding types
• Title format follows the required spec-id · library · pyplots.ai pattern with informative subtitle
• Good use of transparency (opacity=0.4) for overlapping curves as recommended

Weaknesses

• Legend could benefit from a semi-transparent background for better contrast against the curves
• Yellow color (#FFD43B) could be slightly less bright for better visibility on white background
• Could add tooltips to enable interactive exploration of individual observations

Verdict: APPROVED