Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,95 @@
+---
+license: mit
+task_categories:
+  - time-series-forecasting
+language: []
+tags:
+  - air-quality
+  - spatio-temporal
+  - graph-neural-networks
+  - missing-data
+pretty_name: AirQualityBench
+size_categories:
+  - 1G-10G
+---
+
+# AirQualityBench
+
+A global-scale air quality forecasting benchmark featuring **3,720 monitoring stations** across the world with **authentic missing patterns** and physical-scale evaluation.
+
+## Dataset Summary
+
+AirQualityBench provides a realistic testbed for air quality prediction by preserving raw observational characteristics that prior benchmarks have removed through imputation. The dataset spans **5 years (2021–2025)** of hourly measurements across **6 primary pollutants** (PM2.5, PM10, NO₂, O₃, SO₂, CO).
+
+### Key Features
+
+- **Global scale**: 3,720 active stations spanning 7 continents, sourced from [OpenAQ](https://openaq.org/)
+- **Authentic missingness**: No imputation — evaluation uses boolean masks to score only valid observations
+- **Physical-scale metrics**: All statistics computed in original concentration units (μg/m³ for gases, mg/m³ for CO)
+- **Multi-pollutant**: 6 pollutants evaluated jointly
+
+## Data Structure
+
+Each yearly HDF5 file (`aq_compact_{year}.h5`) contains:
+
+| Dataset       | Shape          | Description                          |
+|--------------|----------------|--------------------------------------|
+| `values`     | `(T, 3720, 6)`  | Hourly concentrations (float32)      |
+| `masks`      | `(T, 3720, 6)`  | Valid observation flags (bool)       |
+| `station_ids`| `(3720,)`       | Station identifiers                  |
+| `params`     | `(6,)`          | Pollutant names                      |
+
+Additional files:
+- `adj_mx_10.pkl` — KNN adjacency matrix (k=10, Haversine distance)
+- `scaler.csv` — Per-pollutant mean/std for normalization
+- `global_quality_stats.csv` — Per-station, per-pollutant coverage rates
+- `selected_nodes_metadata.csv` — Station coordinates and metadata
+- `all_stations_coords.csv` — Raw station coordinates from OpenAQ
+
+## Data Splits
+
+| Split      | Years       | Approx. Hours |
+|-----------|-------------|---------------|
+| Train     | 2021–2023   | ~26,280       |
+| Validation| 2024        | ~8,784        |
+| Test      | 2025        | ~8,760        |
+
+## Usage
+
+### Via `datasets` library
+
+```python
+from datasets import load_dataset
+
+# Load training split (2021–2023)
+ds = load_dataset("xuxing123/AirQualityBench", "train", streaming=True)
+for row in ds.take(5):
+    print(row["values"].shape)   # (3720, 6)
+    print(row["masks"].shape)    # (3720, 6)
+```
+
+### Direct HDF5 access (recommended for training)
+
+```python
+import h5py
+
+with h5py.File("aq_compact_2021.h5", "r") as f:
+    values = f["values"][:]   # (8760, 3720, 6)
+    masks = f["masks"][:]     # (8760, 3720, 6)
+```
+
+## Benchmark Code
+
+Training and evaluation code: [github.com/Star-Learning/AirQualityBench](https://github.com/Star-Learning/AirQualityBench)
+
+## Citation
+
+```bibtex
+@article{airqualitybench2025,
+  title={AirQualityBench: A Global-Scale Air Quality Forecasting Benchmark
+         for Spatio-Temporal Graph Neural Networks},
+  author={...},
+  journal={arXiv preprint arXiv:2605.05854},
+  year={2025}
+}
+```