Dataset for Multilingual Multi-Label Classification of Sustainability-Oriented Organizations

This dataset contains 7,347 records derived from the EcoHubMap directory for research on multilingual multi-label text classification of sustainability-oriented organizations. Each record contains an organization’s website URL (website), the text used as input for classification (text), and a JSON array of assigned sustainability categories (fields). The workbook contains 25 distinct category labels, including renewable energy, waste, water, agriculture, and ecotourism. A total of 1,918 records have multiple labels, and category frequencies vary substantially. The dataset accompanies the study “Sparse supervised models versus generative language models for imbalanced multilingual multi-label classification of sustainability-oriented organizations.”