{"spec_id":"pdp-basic","library":"plotly","language":"python","code":"\"\"\" anyplot.ai\npdp-basic: Partial Dependence Plot\nLibrary: plotly 6.7.0 | Python 3.13.13\nQuality: 96/100 | Updated: 2026-05-15\n\"\"\"\n\nimport os\n\nimport numpy as np\nimport plotly.graph_objects as go\nfrom sklearn.datasets import load_diabetes\nfrom sklearn.ensemble import GradientBoostingRegressor\nfrom sklearn.inspection import partial_dependence\n\n\n# Theme tokens\nTHEME = os.getenv(\"ANYPLOT_THEME\", \"light\")\nPAGE_BG = \"#FAF8F1\" if THEME == \"light\" else \"#1A1A17\"\nELEVATED_BG = \"#FFFDF6\" if THEME == \"light\" else \"#242420\"\nINK = \"#1A1A17\" if THEME == \"light\" else \"#F0EFE8\"\nINK_SOFT = \"#4A4A44\" if THEME == \"light\" else \"#B8B7B0\"\nGRID = \"rgba(26,26,23,0.10)\" if THEME == \"light\" else \"rgba(240,239,232,0.10)\"\n\nBRAND = \"#009E73\"  # Okabe-Ito position 1\n\n# Data - Use diabetes dataset for realistic PDP\ndiabetes = load_diabetes()\nX, y = diabetes.data, diabetes.target\nfeature_names = diabetes.feature_names\n\n# Train model\nnp.random.seed(42)\nmodel = GradientBoostingRegressor(n_estimators=100, max_depth=4, random_state=42)\nmodel.fit(X, y)\n\n# Compute partial dependence for BMI (feature 2) - known to have strong effect\nfeature_idx = 2  # bmi\npdp_result = partial_dependence(model, X, features=[feature_idx], kind=\"average\", grid_resolution=100)\nfeature_values = pdp_result[\"grid_values\"][0]\npd_values = pdp_result[\"average\"][0]\n\n# Compute individual conditional expectation (ICE) for uncertainty visualization\nice_result = partial_dependence(model, X, features=[feature_idx], kind=\"individual\", grid_resolution=100)\nice_lines = ice_result[\"individual\"][0]\n\n# Calculate confidence intervals (mean ± 1 std for variability band)\nice_mean = np.mean(ice_lines, axis=0)\nice_std = np.std(ice_lines, axis=0)\nci_lower = ice_mean - ice_std\nci_upper = ice_mean + ice_std\n\n# Center partial dependence at zero for easier interpretation\npd_centered = pd_values - np.mean(pd_values)\nci_lower_centered = ci_lower - np.mean(pd_values)\nci_upper_centered = ci_upper - np.mean(pd_values)\n\n# Create figure\nfig = go.Figure()\n\n# Add confidence band (±1 std)\nfig.add_trace(\n    go.Scatter(\n        x=np.concatenate([feature_values, feature_values[::-1]]),\n        y=np.concatenate([ci_upper_centered, ci_lower_centered[::-1]]),\n        fill=\"toself\",\n        fillcolor=f\"rgba({int(BRAND[1:3], 16)}, {int(BRAND[3:5], 16)}, {int(BRAND[5:7], 16)}, 0.15)\",\n        line=dict(color=\"rgba(255,255,255,0)\"),\n        name=\"±1 Std Deviation\",\n        showlegend=True,\n        hoverinfo=\"skip\",\n    )\n)\n\n# Add partial dependence line\nfig.add_trace(\n    go.Scatter(\n        x=feature_values, y=pd_centered, mode=\"lines\", line=dict(color=BRAND, width=4), name=\"Partial Dependence\"\n    )\n)\n\n# Add rug plot showing distribution of training data\ny_range = np.max(ci_upper_centered) - np.min(ci_lower_centered)\nrug_y = np.full(len(X), np.min(ci_lower_centered) - 0.08 * y_range)\nfig.add_trace(\n    go.Scatter(\n        x=X[:, feature_idx],\n        y=rug_y,\n        mode=\"markers\",\n        marker=dict(symbol=\"line-ns\", size=16, color=INK_SOFT, opacity=0.5, line=dict(width=2)),\n        name=\"Data Distribution\",\n        hoverinfo=\"skip\",\n    )\n)\n\n# Add zero reference line\nfig.add_hline(y=0, line_dash=\"dash\", line_color=INK_SOFT, line_width=2)\n\n# Layout\nfig.update_layout(\n    title=dict(text=\"pdp-basic · plotly · anyplot.ai\", font=dict(size=28, color=INK)),\n    xaxis=dict(\n        title=dict(text=\"BMI (Body Mass Index, standardized)\", font=dict(size=22, color=INK)),\n        tickfont=dict(size=18, color=INK_SOFT),\n        showgrid=True,\n        gridcolor=GRID,\n        gridwidth=1,\n        linecolor=INK_SOFT,\n        zerolinecolor=INK_SOFT,\n    ),\n    yaxis=dict(\n        title=dict(text=\"Partial Dependence (centered)\", font=dict(size=22, color=INK)),\n        tickfont=dict(size=18, color=INK_SOFT),\n        showgrid=True,\n        gridcolor=GRID,\n        gridwidth=1,\n        linecolor=INK_SOFT,\n        zerolinecolor=INK_SOFT,\n    ),\n    paper_bgcolor=PAGE_BG,\n    plot_bgcolor=PAGE_BG,\n    legend=dict(\n        font=dict(size=16, color=INK_SOFT),\n        x=0.02,\n        y=0.98,\n        xanchor=\"left\",\n        yanchor=\"top\",\n        bgcolor=ELEVATED_BG,\n        bordercolor=INK_SOFT,\n        borderwidth=1,\n    ),\n    margin=dict(l=100, r=80, t=120, b=100),\n)\n\n# Save as PNG (4800 x 2700 px) and HTML\nfig.write_image(f\"plot-{THEME}.png\", width=1600, height=900, scale=3)\nfig.write_html(f\"plot-{THEME}.html\", include_plotlyjs=\"cdn\")\n"}