Advanced Visualization

1/**2 * Example 44: Advanced Visualization3 *4 * The function-style plotting API: plot, scatter, bar, hist and heatmap, plus5 * ready-made model diagnostics (confusion matrix, ROC curve, feature importance,6 * elbow curve, residuals). show({ format: "svg" }) renders the current figure7 * and returns the SVG text, which this example reports by length. See example 258 * for writing SVG files with the Figure class.9 */10 11import { cos, linspace, sin, tensor } from "deepbox/ndarray";12import {13  bar,14  heatmap,15  hist,16  plot,17  plotConfusionMatrix,18  plotElbowCurve,19  plotFeatureImportance,20  plotResiduals,21  plotRocCurve,22  scatter,23  show,24} from "deepbox/plot";25 26console.log("=".repeat(60));27console.log("Example 44: Advanced Visualization");28console.log("=".repeat(60));29 30// ============================================================================31// Part 1: Line Plot32// ============================================================================33console.log("\nPart 1: Line Plot");34console.log("-".repeat(60));35 36// 100 x values from 0 to 2 pi37const x = linspace(0, 2 * Math.PI, 100);38const ySin = sin(x);39const yCos = cos(x);40 41// Plot sine and cosine curves42plot(x, ySin, { color: "blue", label: "sin(x)" });43plot(x, yCos, { color: "red", label: "cos(x)" });44const lineSvg = show({ format: "svg" });45console.log("Line plot (sin/cos) rendered:");46console.log(`  SVG output: ${lineSvg.svg.length} characters`);47 48// ============================================================================49// Part 2: Scatter Plot50// ============================================================================51console.log("\nPart 2: Scatter Plot");52console.log("-".repeat(60));53 54const scatterX = tensor([1, 2, 3, 4, 5, 6, 7, 8, 9, 10]);55const scatterY = tensor([2.1, 3.9, 6.2, 7.8, 10.1, 12.0, 13.8, 16.1, 18.2, 19.9]);56 57scatter(scatterX, scatterY, { color: "blue", label: "data points" });58const scatterSvg = show({ format: "svg" });59console.log("Scatter plot rendered:");60console.log(`  SVG output: ${scatterSvg.svg.length} characters`);61 62// ============================================================================63// Part 3: Bar Chart64// ============================================================================65console.log("\nPart 3: Bar Chart");66console.log("-".repeat(60));67 68const categories = tensor([1, 2, 3, 4, 5]);69const values = tensor([23, 45, 12, 67, 34]);70 71bar(categories, values, { color: "steelblue", label: "Sales" });72const barSvg = show({ format: "svg" });73console.log("Bar chart rendered:");74console.log(`  SVG output: ${barSvg.svg.length} characters`);75 76// ============================================================================77// Part 4: Histogram78// ============================================================================79console.log("\nPart 4: Histogram");80console.log("-".repeat(60));81 82// Deterministic pseudo-random values83const histData: number[] = [];84for (let i = 0; i < 200; i++) {85  // Box-Muller transform: two uniform values give one normal value86  const u1 = (i + 1) / 201;87  const u2 = (((i * 7 + 3) % 200) + 1) / 201;88  histData.push(Math.sqrt(-2 * Math.log(u1)) * Math.cos(2 * Math.PI * u2));89}90 91hist(tensor(histData), 20, { color: "green", label: "Normal-like data" });92const histSvg = show({ format: "svg" });93console.log("Histogram (20 bins) rendered:");94console.log(`  SVG output: ${histSvg.svg.length} characters`);95 96// ============================================================================97// Part 5: Heatmap98// ============================================================================99console.log("\nPart 5: Heatmap");100console.log("-".repeat(60));101 102const heatmapData = tensor([103  [1, 2, 3, 4],104  [5, 6, 7, 8],105  [9, 10, 11, 12],106  [13, 14, 15, 16],107]);108 109heatmap(heatmapData, { label: "4x4 heatmap" });110const heatSvg = show({ format: "svg" });111console.log("Heatmap rendered:");112console.log(`  SVG output: ${heatSvg.svg.length} characters`);113 114// ============================================================================115// Part 6: Confusion Matrix116// ============================================================================117console.log("\nPart 6: Confusion Matrix");118console.log("-".repeat(60));119 120const confMatrix = tensor([121  [45, 5, 2],122  [3, 40, 7],123  [1, 4, 43],124]);125 126plotConfusionMatrix(confMatrix, ["Cat", "Dog", "Bird"]);127const cmSvg = show({ format: "svg" });128console.log("Confusion Matrix rendered:");129console.log(`  SVG output: ${cmSvg.svg.length} characters`);130console.log("  Classes: Cat, Dog, Bird");131 132// ============================================================================133// Part 7: ROC Curve134// ============================================================================135console.log("\nPart 7: ROC Curve");136console.log("-".repeat(60));137 138// Example ROC curve points139const fpr = tensor([0, 0.05, 0.1, 0.2, 0.3, 0.5, 0.7, 1.0]);140const tpr = tensor([0, 0.4, 0.65, 0.8, 0.88, 0.94, 0.98, 1.0]);141 142plotRocCurve(fpr, tpr, 0.87);143const rocSvg = show({ format: "svg" });144console.log("ROC Curve rendered (AUC passed in: 0.87):");145console.log(`  SVG output: ${rocSvg.svg.length} characters`);146 147// ============================================================================148// Part 8: Feature Importance149// ============================================================================150console.log("\nPart 8: Feature Importance");151console.log("-".repeat(60));152 153const importances = tensor([0.35, 0.25, 0.15, 0.12, 0.08, 0.05]);154const featureNames = ["income", "age", "credit_score", "tenure", "balance", "products"];155 156plotFeatureImportance(importances, featureNames);157const fiSvg = show({ format: "svg" });158console.log("Feature Importance chart rendered:");159console.log(`  SVG output: ${fiSvg.svg.length} characters`);160 161// ============================================================================162// Part 9: Elbow Curve163// ============================================================================164console.log("\nPart 9: Elbow Curve (KMeans)");165console.log("-".repeat(60));166 167const kValues = tensor([2, 3, 4, 5, 6, 7, 8]);168const inertias = tensor([500, 300, 180, 120, 100, 90, 85]);169 170plotElbowCurve(kValues, inertias);171const elbowSvg = show({ format: "svg" });172console.log("Elbow Curve rendered:");173console.log(`  SVG output: ${elbowSvg.svg.length} characters`);174console.log("  The inertia stops dropping quickly at k = 4 or 5, the elbow");175 176// ============================================================================177// Part 10: Residual Plot178// ============================================================================179console.log("\nPart 10: Residual Plot");180console.log("-".repeat(60));181 182const yTrue = tensor([3, 5, 7, 9, 11, 13, 15]);183const yPred = tensor([3.1, 4.8, 7.3, 8.7, 11.2, 12.8, 15.1]);184 185plotResiduals(yTrue, yPred);186const resSvg = show({ format: "svg" });187console.log("Residual Plot rendered:");188console.log(`  SVG output: ${resSvg.svg.length} characters`);189 190// ============================================================================191// Summary192// ============================================================================193console.log("\nKey Takeaways");194console.log("-".repeat(60));195console.log("• plot, scatter: line and point plots");196console.log("• bar: bar chart for categories");197console.log("• hist: histogram of one variable");198console.log("• heatmap: color-coded matrix");199console.log("• plotConfusionMatrix: classification results per class");200console.log("• plotRocCurve: true positive rate against false positive rate");201console.log("• plotFeatureImportance: which features matter most");202console.log("• plotElbowCurve: inertia against k, to choose the number of clusters");203console.log("• plotResiduals: residuals of a regression model");204console.log(205  '• show({ format: "svg" }) returns SVG text. Figure.renderPNG() renders PNG in Node.js.'206);207 208console.log("\nAdvanced Visualization Example Complete!");209console.log("=".repeat(60));