UTU-House Dataset

The UTU-House Dataset is the first Indonesian house image dataset labeled for socioeconomic-based classification (Simple, Middle, Luxury). It was developed to support automated, objective alternatives to manual house type assessment used in Indonesian policy contexts such as tuition fee determination, scholarship eligibility, and social assistance programs. - Total images: 3,000 RGB images - Classes: Simple, Middle, Luxury (1,000 images per class, balanced) - Geographic coverage: 31 Indonesian provinces - Image sources: (1) student-contributed uploads from Universitas Teuku Umar (UTU), and (2) manually captured screenshots from Google Street View - Format: JPG, 224×224 Classification Criteria Class labels were assigned based on three visual indicators correlated with housing value under Indonesian Government Regulation No. 12/2021 and BPS (Statistics Indonesia) housing quality indicators: Dataset Structure UTU-House_Dataset.zip ├── README.md ├── metadata.csv ├── simple/ │ ├── simple_0001.jpg │ ├── simple_0002.jpg │ └── ... (1,000 images) ├── middle/ │ ├── middle_0001.jpg │ └── ... (1,000 images) └── luxury/ ├── luxury_0001.jpg └── ... (1,000 images) If you use this dataset, please cite: Ainuddin, Nizamuddin& Safwan. (2026). UTU-House Dataset [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.21294955 Contact For questions, access requests, or reporting issues with the dataset, please contact: - Ainuddin — [e-mail] - Affiliation: Universitas Syiah Kuala (research) / Universitas Teuku Umar (data provision)