feat(altair): implement chernoff-basic

[github-actions]

Implementation: chernoff-basic - altair

Implements the altair version of chernoff-basic.

File: plots/chernoff-basic/implementations/altair.py

Parent Issue: #3003

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

The plot displays 12 Chernoff faces arranged in a 4x3 grid on a white background. Each face is constructed from layered circular points representing facial features. The faces are colored in three distinct colors based on species: blue (#306998) for setosa, yellow (#FFD43B) for versicolor, and light blue (#4B8BBE) for virginica. Each face has two dark circular eyes with white pupil highlights, a small gray nose, and a red circular mouth. The faces vary in size (face width and height) and feature proportions (eye size, mouth width) based on the Iris dataset measurements. Sample labels (Sample 1 through Sample 12) appear below each face. A feature mapping legend in the lower left explains: Face width = sepal length, Face height = sepal width, Eye size = petal length, Mouth width = petal width. A species color legend appears in the lower right. The title "chernoff-basic · altair · pyplots.ai" is at the top with a subtitle explaining the visualization.

Quality Score: 78/100

Criteria Checklist

Visual Quality (31/40 pts)

• VQ-01: Text Legibility (7/10) - Title and labels are readable but subtitle and mapping text are relatively small
• VQ-02: No Overlap (8/8) - No overlapping text, all labels clearly separated
• VQ-03: Element Visibility (6/8) - Faces are visible but some are quite small, making feature differences harder to distinguish
• VQ-04: Color Accessibility (4/5) - Three colors distinguishable; blue shades for setosa and virginica are somewhat similar
• VQ-05: Layout Balance (3/5) - Faces occupy central area but significant whitespace at top and sides; mapping legend overlaps with Sample 9 label
• VQ-06: Axis Labels (2/2) - N/A for this plot type (no axes), but feature mapping provided
• VQ-07: Grid & Legend (1/2) - Species legend present and clear; mapping legend placement is awkward (overlaps with face label)

Spec Compliance (22/25 pts)

• SC-01: Plot Type (8/8) - Correct Chernoff faces visualization
• SC-02: Data Mapping (5/5) - Variables correctly mapped to facial features (sepal length→width, sepal width→height, etc.)
• SC-03: Required Features (4/5) - Has face components, grid layout, color by category; lacks eyebrow feature mentioned in spec
• SC-04: Data Range (3/3) - All 12 samples displayed with normalized data
• SC-05: Legend Accuracy (0/2) - Species legend colors are inconsistent with actual face colors displayed
• SC-06: Title Format (2/2) - Uses correct format: "chernoff-basic · altair · pyplots.ai"

Data Quality (17/20 pts)

• DQ-01: Feature Coverage (7/8) - Shows variation across species with different face sizes and features; good range of values
• DQ-02: Realistic Context (7/7) - Uses Iris dataset, a classic multivariate dataset perfect for this visualization
• DQ-03: Appropriate Scale (3/5) - Normalized 0-1 values as specified; some faces appear too similar in size

Code Quality (8/10 pts)

• CQ-01: KISS Structure (3/3) - Linear flow: imports → data → face construction → chart layers → save
• CQ-02: Reproducibility (3/3) - Uses np.random.seed(42)
• CQ-03: Clean Imports (2/2) - Only necessary imports (altair, numpy, pandas)
• CQ-04: No Deprecated API (0/1) - Uses order encoding which may cause rendering issues
• CQ-05: Output Correct (0/1) - Saves both plot.png and plot.html (correct)

Library Features (0/5 pts)

• LF-01: Uses distinctive library features (0/5) - Implementation uses mark_point for everything rather than leveraging Altair's declarative grammar for layered compositions; essentially simulates faces with sized points rather than using Altair's strengths

Strengths

• Creative approach to implementing Chernoff faces in Altair using layered point marks
• Good use of Iris dataset as a realistic multivariate example
• Clear feature-to-facial-attribute mapping documented in the visualization
• Species color coding allows pattern recognition across groups
• Proper grid layout for comparing 12 faces side by side

Weaknesses

• Feature mapping legend text overlaps with Sample 9 label area - needs repositioning
• Legend color points don't match actual face colors (setosa legend shows small blue dot but faces use larger translucent blue)
• Faces lack eyebrow feature mentioned in specification (eyebrow slant)
• All facial features are circles - mouth should be more elliptical/curved for better face representation
• Small faces make subtle feature differences hard to distinguish visually
• Implementation doesn't leverage Altair's declarative strengths - uses manual point placement instead of proper layered mark compositions

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 displays 12 Chernoff faces arranged in a 3x4 grid on a white background. Each face is composed of circular shapes with facial features including eyes (black circles with white pupils), eyebrows (dark dots), a small nose, and a red curved mouth. The faces are colored by species: blue (#306998) for setosa, yellow (#FFD43B) for versicolor, and light blue (#4B8BBE) for virginica. A legend in the top-right shows the three species colors. The bottom-left contains a feature mapping key explaining how variables map to facial features. Each face is labeled below with "Sample 1" through "Sample 12". The title "chernoff-basic · altair · pyplots.ai" appears at the top with a descriptive subtitle.

Quality Score: 85/100

Criteria Checklist

Visual Quality (32/40 pts)

• VQ-01: Text Legibility (8/10) - Title and labels readable, but could be larger
• VQ-02: No Overlap (8/8) - No overlapping elements
• VQ-03: Element Visibility (6/8) - Faces visible but feature variations subtle
• VQ-04: Color Accessibility (5/5) - Blue/yellow/light blue distinguishable
• VQ-05: Layout Balance (3/5) - Upper portion has significant empty space
• VQ-06: Axis Labels (N/A) - Not applicable for this chart type
• VQ-07: Grid & Legend (2/2) - Legend well placed, no grid needed

Spec Compliance (23/25 pts)

• SC-01: Plot Type (8/8) - Correct Chernoff faces visualization
• SC-02: Data Mapping (5/5) - Variables mapped to facial features correctly
• SC-03: Required Features (4/5) - Most features present, face shape variation subtle
• SC-04: Data Range (3/3) - All 12 samples visible
• SC-05: Legend Accuracy (2/2) - Species legend accurate
• SC-06: Title Format (2/2) - Uses "{spec-id} · {library} · pyplots.ai"

Data Quality (18/20 pts)

• DQ-01: Feature Coverage (6/8) - Shows variation but facial differences subtle
• DQ-02: Realistic Context (7/7) - Iris dataset is appropriate and neutral
• DQ-03: Appropriate Scale (5/5) - Normalized 0-1 values as specified

Code Quality (9/10 pts)

• CQ-01: KISS Structure (3/3) - Linear flow, no functions/classes
• CQ-02: Reproducibility (3/3) - np.random.seed(42) present
• CQ-03: Clean Imports (2/2) - Only altair, numpy, pandas
• CQ-04: No Deprecated API (1/1) - Current Altair API
• CQ-05: Output Correct (0/1) - Saves both PNG and HTML (correct)

Library Features (3/5 pts)

• LF-01: Uses distinctive library features (3/5) - Creative use of layered marks, but could leverage more Altair-specific features like selection or faceting

Strengths

• Creative use of Altair layered marks to construct face components from points
• Proper feature mapping legend explaining variable-to-facial-feature correspondence
• Good color coding by species with clear legend
• Correct title format following pyplots.ai conventions
• Reproducible with fixed random seed

Weaknesses

• Face shapes are circular (all same shape) - should vary face width/height more visibly based on the encoded variables
• Faces appear somewhat small relative to canvas with significant empty space in upper portion
• Sample labels could be more descriptive (e.g., include species name)

Verdict: REJECTED

The implementation demonstrates a creative approach to Chernoff faces in Altair using layered point marks. However, the facial feature variations (especially face width/height) are too subtle to effectively distinguish between samples, which is the core purpose of Chernoff faces. The layout also has too much empty space at the top. Score of 85 does not meet the 90+ threshold for immediate approval.

[github-actions]

🔧 Repair Attempt 2/3

Applied fixes based on AI review feedback.

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

---

🤖 impl-repair

[claude]

AI Review - Attempt 3/3

Image Description

The plot displays 12 Chernoff faces arranged in a 4-column by 3-row grid on a white background. Each face is composed of overlapping circles creating an oval face shape, with dark eyes (with white pupil highlights), gray eyebrows, a gray nose, and a red mouth. Faces are colored by species: setosa faces are blue (#306998), versicolor faces are yellow (#FFD43B), and virginica faces are light blue (#4B8BBE). The faces visibly vary in size and proportions - setosa faces tend to be narrower with smaller eyes/mouths, while virginica faces are larger with bigger features. A legend in the upper right shows the three species with colored circles. The bottom left contains a "Feature Mapping" key explaining which Iris measurements map to which facial features. Each face has a label below it (e.g., "Sample 1 (setosa)"). The title "chernoff-basic · altair · pyplots.ai" appears at the top with a descriptive subtitle.

Quality Score: 87/100

Criteria Checklist

Visual Quality (34/40 pts)

• VQ-01: Text Legibility (9/10) - Title and labels are readable, subtitle and feature mapping text could be slightly larger
• VQ-02: No Overlap (8/8) - No overlapping text elements
• VQ-03: Element Visibility (6/8) - Faces are visible but the circle-based approximation of face shapes looks somewhat "blobby" - the overlapping circles create uneven edges
• VQ-04: Color Accessibility (5/5) - Blue, yellow, and light blue are distinguishable for colorblind users
• VQ-05: Layout Balance (4/5) - Good grid layout utilizing most of the canvas, though the feature mapping overlaps slightly with Sample 9's label
• VQ-06: Axis Labels (0/2) - N/A for this plot type (no axes), but feature mapping serves similar purpose - 1/2
• VQ-07: Grid & Legend (2/2) - Legend is well placed and clear

Spec Compliance (23/25 pts)

• SC-01: Plot Type (8/8) - Correct Chernoff faces visualization
• SC-02: Data Mapping (5/5) - Variables correctly mapped to facial features (sepal_length→width, sepal_width→height, petal_length→eye_size, petal_width→mouth_width)
• SC-03: Required Features (4/5) - Has most features: face width/height, eye size, mouth width, eyebrow slant; grid layout present; color by group present
• SC-04: Data Range (3/3) - All 12 samples visible
• SC-05: Legend Accuracy (2/2) - Species legend is correct
• SC-06: Title Format (2/2) - Uses correct format "chernoff-basic · altair · pyplots.ai"

Data Quality (18/20 pts)

• DQ-01: Feature Coverage (7/8) - Shows good variation across samples; setosa vs virginica clearly distinguishable; could show more extreme variations
• DQ-02: Realistic Context (7/7) - Iris dataset is a classic, appropriate example for multivariate visualization
• DQ-03: Appropriate Scale (4/5) - Normalized 0-1 values as specified; some faces look similar in the versicolor group

Code Quality (9/10 pts)

• CQ-01: KISS Structure (3/3) - Linear structure with imports → data → computation → chart
• CQ-02: Reproducibility (3/3) - Uses np.random.seed(42)
• CQ-03: Clean Imports (2/2) - Only necessary imports (altair, numpy, pandas)
• CQ-04: No Deprecated API (1/1) - Uses current Altair 6.0 API
• CQ-05: Output Correct (0/1) - Saves both plot.png and plot.html correctly

Library Features (3/5 pts)

• LF-01: Uses distinctive library features (3/5) - Uses Altair's layered chart composition and encoding system effectively; however, Altair doesn't have native ellipse marks, so the circle-based approach is a reasonable workaround, though not leveraging Altair's declarative strengths as much as possible

Strengths

• Creative approach to implement Chernoff faces in Altair using layered point marks to simulate face shapes
• Good use of color encoding by species for immediate visual grouping
• Well-organized grid layout with 12 faces showing clear variation
• Feature mapping legend is helpful for understanding the visualization
• Proper title format and informative subtitle
• Clean, reproducible code with appropriate seed

Weaknesses

• The circle-based face approximation creates "blobby" irregular face shapes rather than smooth ellipses
• Feature mapping text in bottom left slightly overlaps with Sample 9's label area
• Some versicolor faces appear quite similar, making individual differentiation harder

Verdict: APPROVED

This is attempt 3/3. The implementation demonstrates a creative solution to a challenging visualization type that Altair doesn't natively support. While the circle-based face shapes are not as smooth as true ellipses, the visualization effectively conveys the multivariate data through facial feature variations. The faces are distinguishable, the color grouping is clear, and the overall presentation is professional. Score of 87 exceeds the 50-point threshold for final attempt approval.

[github-actions]

🔧 Repair Attempt 3/3

Applied fixes based on AI review feedback.

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

---

🤖 impl-repair

[claude]

AI Review - Attempt 3/3

Image Description

The plot displays 12 Chernoff faces arranged in a 4×3 grid layout. Each face represents an Iris dataset sample, with faces colored by species: blue (#306998) for setosa, yellow (#FFD43B) for versicolor, and light blue (#4B8BBE) for virginica. Each face has distinct features - eyes (dark circles with white highlights), eyebrows (small dark dots), a nose (gray dot), and a red mouth. Face dimensions vary based on measurements: setosa samples (1, 4, 7, 11) have smaller/narrower faces with tiny eyes; virginica samples (2, 5, 8, 10) have larger faces with prominent eyes; versicolor samples (3, 6, 9, 12) fall in between. A species legend is positioned in the top right corner. Feature mapping explanation is in the bottom left. Each face is labeled below with sample number and species name.

Quality Score: 87/100

Criteria Checklist

Visual Quality (34/40 pts)

• VQ-01: Text Legibility (9/10) - Title, subtitle, and labels are readable; feature mapping text is slightly small
• VQ-02: No Overlap (4/8) - Feature mapping text overlaps with "Sample 9 (versicolor)" label in bottom left
• VQ-03: Element Visibility (8/8) - Faces and facial features are clearly visible and distinguishable
• VQ-04: Color Accessibility (5/5) - Blue/yellow/light-blue palette is colorblind-friendly
• VQ-05: Layout Balance (5/5) - Good 4×3 grid layout, faces well-spaced, good canvas utilization
• VQ-06: Axis Labels (2/2) - N/A for Chernoff faces, feature mapping serves this purpose
• VQ-07: Grid & Legend (1/2) - Legend present and well-placed; feature mapping serves as guide

Spec Compliance (25/25 pts)

• SC-01: Plot Type (8/8) - Correct Chernoff faces implementation
• SC-02: Data Mapping (5/5) - Variables correctly mapped to facial features (width, height, eyes, mouth, eyebrows)
• SC-03: Required Features (5/5) - All spec features present: multiple variables, grid layout, color by group
• SC-04: Data Range (3/3) - Normalized 0-1 values properly represented
• SC-05: Legend Accuracy (2/2) - Species legend correctly labeled
• SC-06: Title Format (2/2) - Correct format: "chernoff-basic · altair · pyplots.ai"

Data Quality (18/20 pts)

• DQ-01: Feature Coverage (7/8) - Shows variation across species; some faces could show more pronounced differences
• DQ-02: Realistic Context (7/7) - Iris dataset is a classic, neutral, scientific example
• DQ-03: Appropriate Scale (4/5) - Normalized values appropriate; feature variation visible but moderate

Code Quality (10/10 pts)

• CQ-01: KISS Structure (3/3) - Linear flow: imports → data → face construction → chart → save
• CQ-02: Reproducibility (3/3) - Uses np.random.seed(42)
• CQ-03: Clean Imports (2/2) - Only necessary imports (altair, numpy, pandas)
• CQ-04: No Deprecated API (1/1) - Current Altair 5.x API
• CQ-05: Output Correct (1/1) - Saves both plot.png and plot.html

Library Features (0/5 pts)

• LF-01: Uses distinctive library features (0/5) - Creative use of mark_point to build faces, but fundamentally using basic scatter functionality. Altair's declarative grammar not leveraged for interactivity or linked selections that would showcase the library.

Strengths

• Creative implementation of Chernoff faces using Altair's mark_point with layered scatter points
• Excellent color coding by species with colorblind-friendly palette
• Well-organized grid layout with good canvas utilization
• Feature mapping legend clearly explains variable-to-feature correspondence
• Good use of Iris dataset as a classic multivariate example

Weaknesses

• Feature mapping text overlaps with "Sample 9 (versicolor)" label causing text collision
• Does not leverage Altair's distinctive features (interactivity, tooltips, selections)
• Face feature variations could be more visually pronounced between samples

Verdict: APPROVED

Rationale: This is attempt 3/3 and the implementation scores 87/100 (≥50 threshold). While the text overlap issue persists, the overall implementation successfully demonstrates Chernoff faces with creative use of Altair's mark_point. The face construction is sophisticated, the species coloring is effective, and the layout is well-organized. The implementation meets the quality bar for inclusion in the repository.

[github-actions]

AI Review - Final Status

Score: 87/100 (Acceptable)

After 3 repair attempts, altair reached 87/100.

Score ≥ 50 is acceptable for the repository. Merging.

---

🤖 impl-review