feat(altair): implement gain-curve (#2458)

github-actions[bot] · web-flow · commit 88ff6aab4120 · 2025-12-29T22:55:41.000Z
## Implementation: `gain-curve` - altair Implements the **altair** version of `gain-curve`. **File:** `plots/gain-curve/implementations/altair.py` --- :robot: *[impl-generate workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20584329173)* --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
diff --git a/plots/gain-curve/implementations/altair.py b/plots/gain-curve/implementations/altair.py
@@ -0,0 +1,137 @@
+""" pyplots.ai
+gain-curve: Cumulative Gains Chart
+Library: altair 6.0.0 | Python 3.13.11
+Quality: 91/100 | Created: 2025-12-29
+"""
+
+import altair as alt
+import numpy as np
+import pandas as pd
+
+
+# Data - Simulated model predictions for customer response
+np.random.seed(42)
+n_samples = 1000
+
+# Generate realistic customer response data
+# Model with good discrimination - scores correlate with actual outcomes
+base_score = np.random.beta(2, 5, n_samples)  # Skewed distribution of base scores
+noise = np.random.normal(0, 0.15, n_samples)
+y_score = np.clip(base_score + noise, 0, 1)
+
+# Generate actual outcomes with correlation to scores
+response_prob = 0.3 * y_score + 0.05  # Higher score = higher probability of positive
+y_true = (np.random.random(n_samples) < response_prob).astype(int)
+
+# Calculate cumulative gains curve
+sorted_indices = np.argsort(y_score)[::-1]  # Sort by score descending
+y_true_sorted = y_true[sorted_indices]
+
+# Calculate cumulative percentages
+cumulative_positives = np.cumsum(y_true_sorted)
+total_positives = y_true_sorted.sum()
+pct_population = np.arange(1, n_samples + 1) / n_samples * 100
+pct_gain = cumulative_positives / total_positives * 100
+
+# Add origin point for complete curve
+pct_population = np.insert(pct_population, 0, 0)
+pct_gain = np.insert(pct_gain, 0, 0)
+
+# Subsample for smoother visual (keep every 10th point plus endpoints)
+sample_idx = np.concatenate([[0], np.arange(10, len(pct_population) - 1, 10), [len(pct_population) - 1]])
+pct_population_smooth = pct_population[sample_idx]
+pct_gain_smooth = pct_gain[sample_idx]
+
+# Create DataFrame for gains curve
+df_gains = pd.DataFrame({"population": pct_population_smooth, "gain": pct_gain_smooth, "Type": "Model"})
+
+# Create diagonal reference line (random selection baseline)
+df_baseline = pd.DataFrame({"population": [0, 100], "gain": [0, 100], "Type": "Random (Baseline)"})
+
+# Combine data
+df_combined = pd.concat([df_gains, df_baseline], ignore_index=True)
+
+# Create gain curve chart with proper styling
+model_curve = (
+    alt.Chart(df_combined[df_combined["Type"] == "Model"])
+    .mark_line(
+        strokeWidth=4,
+        color="#306998",
+        interpolate="monotone",  # Smooth interpolation
+    )
+    .encode(
+        x=alt.X(
+            "population:Q",
+            title="Population Targeted (%)",
+            scale=alt.Scale(domain=[0, 100]),
+            axis=alt.Axis(titleFontSize=22, labelFontSize=18, tickCount=10),
+        ),
+        y=alt.Y(
+            "gain:Q",
+            title="Positive Cases Captured (%)",
+            scale=alt.Scale(domain=[0, 100]),
+            axis=alt.Axis(titleFontSize=22, labelFontSize=18, tickCount=10),
+        ),
+    )
+)
+
+# Add shaded area under model curve
+area_model = (
+    alt.Chart(df_combined[df_combined["Type"] == "Model"])
+    .mark_area(opacity=0.15, color="#306998", interpolate="monotone")
+    .encode(x="population:Q", y="gain:Q")
+)
+
+# Baseline diagonal line
+baseline_line = (
+    alt.Chart(df_combined[df_combined["Type"] == "Random (Baseline)"])
+    .mark_line(strokeWidth=3, strokeDash=[8, 4], color="#888888")
+    .encode(x="population:Q", y="gain:Q")
+)
+
+# Custom legend using text and lines
+legend_model_line = (
+    alt.Chart(pd.DataFrame({"x": [8, 18], "y": [94, 94]}))
+    .mark_line(strokeWidth=4, color="#306998")
+    .encode(x=alt.X("x:Q"), y=alt.Y("y:Q"))
+)
+
+legend_model_text = (
+    alt.Chart(pd.DataFrame({"x": [20], "y": [94], "text": ["Model"]}))
+    .mark_text(align="left", fontSize=18, color="#333333")
+    .encode(x="x:Q", y="y:Q", text="text:N")
+)
+
+legend_baseline_line = (
+    alt.Chart(pd.DataFrame({"x": [8, 18], "y": [88, 88]}))
+    .mark_line(strokeWidth=3, strokeDash=[8, 4], color="#888888")
+    .encode(x=alt.X("x:Q"), y=alt.Y("y:Q"))
+)
+
+legend_baseline_text = (
+    alt.Chart(pd.DataFrame({"x": [20], "y": [88], "text": ["Random (Baseline)"]}))
+    .mark_text(align="left", fontSize=18, color="#333333")
+    .encode(x="x:Q", y="y:Q", text="text:N")
+)
+
+# Combine all layers
+chart = (
+    alt.layer(
+        area_model,
+        baseline_line,
+        model_curve,
+        legend_model_line,
+        legend_model_text,
+        legend_baseline_line,
+        legend_baseline_text,
+    )
+    .properties(
+        width=1600, height=900, title=alt.Title("gain-curve · altair · pyplots.ai", fontSize=28, anchor="middle")
+    )
+    .configure_axis(grid=True, gridOpacity=0.3, gridDash=[2, 2])
+    .configure_view(strokeWidth=0)
+)
+
+# Save as PNG and HTML
+chart.save("plot.png", scale_factor=3.0)
+chart.save("plot.html")
diff --git a/plots/gain-curve/metadata/altair.yaml b/plots/gain-curve/metadata/altair.yaml
@@ -0,0 +1,27 @@
+library: altair
+specification_id: gain-curve
+created: '2025-12-29T22:47:44Z'
+updated: '2025-12-29T22:55:18Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20584329173
+issue: 0
+python_version: 3.13.11
+library_version: 6.0.0
+preview_url: https://storage.googleapis.com/pyplots-images/plots/gain-curve/altair/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/gain-curve/altair/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/gain-curve/altair/plot.html
+quality_score: 91
+review:
+  strengths:
+  - Excellent use of Altair's layered composition pattern with separate charts for
+    model curve, area fill, baseline, and custom legend elements
+  - Clean implementation of monotone interpolation for smooth curve rendering
+  - Good visual hierarchy with shaded area under the model curve providing clear differentiation
+    from baseline
+  - Proper use of Altair's configure_axis and configure_view for styling
+  - Custom legend implementation shows creative use of mark_line and mark_text layers
+  weaknesses:
+  - Custom legend placement in upper-left overlaps the plot data area slightly; consider
+    moving to a less intrusive location
+  - The axis labels include (%) in the title text but this could be formatted more
+    consistently
