This project involves implementing the K-Means clustering algorithm from scratch to analyze Airbnb listing data in Greater Manchester from June 2024. It aims to explore the geographical distribution, price stratification, and property type differences of listings, providing data support for market positioning and pricing strategies.
- Self-developed Clustering Algorithm: Handcrafted the K-Means algorithm, including Euclidean distance calculation, random centroid initialization, iterative updates, and convergence criteria, without relying on external machine learning libraries.
- Multi-dimensional Data Analysis: Constructed three clustering models: Model 1: Reveals geographical distribution based on latitude and longitude. Model 2: Differentiates budget and luxury listings by incorporating price data. Model 3: Identifies spatial distribution of various accommodation types by integrating property type information.
- Data Preprocessing & Visualization: Cleaned raw data, performed feature engineering, handled outliers, and visualized clustering results and market segmentation using charts like scatter plots and box plots.