ArchEGraph/ArchEGraph-demo

ArchEGraph-demo是一个用于建筑级别能源建模的多模态建筑模拟数据集。数据集链接了四种模态：建筑级别的描述符（在`building/`中，NPZ格式）、天气特征（在`weather/`中，NPZ格式）、几何/拓扑特征（在`geometry/`中，NPZ格式）和每小时的能量信号（在`energy/`中，NPZ格式）。一个清单文件（`manifest.csv`）将每个样本映射到其对应的天气/建筑源和能量文件。数据集包含300个样本，75个独特的建筑和48个独特的天气ID。数据集的用途包括建筑能源预测和模拟替代建模、跨天气、建筑和几何数据的多模态学习、领域适应和迁移学习、城市建筑库存分析的表征学习等。ArchEGraph-demo is a multimodal building simulation dataset for building-level energy modeling. The dataset incorporates four modalities: building-level descriptors (stored in `building/` in NPZ format), weather features (stored in `weather/` in NPZ format), geometric/topological features (stored in `geometry/` in NPZ format), and hourly energy signals (stored in `energy/` in NPZ format). A manifest file (`manifest.csv`) maps each sample to its corresponding weather, building source and energy files. The dataset contains 300 samples, 75 unique buildings and 48 unique weather IDs. Its applicable scenarios include building energy prediction and surrogate modeling for building simulations, multimodal learning across weather, building and geometric data, domain adaptation and transfer learning, as well as representation learning for urban building stock analysis, etc.