Hierarchical Data Analysis: Agglomerative Clustering & Tree Map Visualizations
In this video, we dive deep into the methods and techniques behind Hierarchical Data Analysis. Whether you are looking to understand how to cluster complex datasets or how to visualize hierarchical structures effectively, this lecture covers the essential algorithms and design strategies.
Topics Covered:
• Hierarchical Clustering Fundamentals: Understanding the difference between Agglomerative (bottom-up) and Divisive (top-down) approaches.
• Clustering Algorithms: A step-by-step look at the Agglomerative Clustering algorithm, including computing distance matrices and merging clusters.
• Distance Metrics: We compare different proximity definitions including Single-link, Complete-link, Average-link, and Ward’s method, analyzing their strengths and sensitivity to noise/outliers.
• Node-Link Visualizations: Exploring traditional tree metaphors, rooted trees, and advanced layouts like Cone Trees (3D) and Hyperbolic Trees.
• Space-Filling Visualizations: A comprehensive guide to Tree Maps, covering specific algorithms (Slice-and-dice, Squarified, Strip, and Pivot-by-size) and metrics for evaluating aspect ratios.
• Advanced Techniques: Visualizing structure with Cushion Treemaps, Voronoi Treemaps (Fortune's Algorithm), and Circle Packing.
Key Concepts: Dendrograms, Euclidean Distance, Aspect Ratios, Space Trees, and Hierarchical Structures.
#DataScience #DataVisualization #Clustering #MachineLearning #Treemaps #HierarchicalData #Algorithms