This project focuses on analyzing and classifying dwelling types to better understand and address housing poverty issues in Korea. By leveraging machine learning models and data analytics, the project aims to provide actionable insights for policymakers and researchers working to alleviate housing challenges. This project analyzes the 2020 study "Dwelling Type Classification for Disaster Risk Assessment Using Satellite Imagery" and redesigns it to fit the context of Korea. This project is currently in progress.
Understand Housing Poverty: Identify key indicators of housing poverty in Korea. Classify Dwelling Types: Develop a classification system for various dwelling types based on collected data. Policy Recommendations: Provide data-driven insights to support housing policy improvements.
- Data Collection and Preprocessing: Sourced housing-related data from various public and private databases. Cleaned, normalized, and structured data for analysis.
- Exploratory Data Analysis (EDA): Investigated patterns and correlations within the dataset. Identified key features for classification and prediction.
- Model Development: Implemented machine learning algorithms to classify dwelling types. Evaluated model performance using metrics such as accuracy, precision, recall, and F1 score.
- Findings and Insights: Generated visualizations and summaries to highlight critical insights. Developed actionable recommendations for addressing housing poverty.
Classification Models: Test various machine learning models, including Random Forest and Gradient Boosting. Feature Importance: Identify influential features contributing to housing classification. Ethical Considerations: Ensure compliance with ethical standards in handling and analyzing housing data.