DiTEC-WDN: 36 Water Distribution Networks with 1,000 Scenarios Each

36 Water Distribution Networks (WDNs) each contain 1,000 distinct scenarios. Each scenario is a timeseries of directed shared-topology graphs, representing a spatiotemporal graph where nodes and edges are multivariate time series. The dataset was created by author 'rugds' and was last updated on 2025-11-25.

Use Cases

• Train spatiotemporal graph neural networks based on the timeseries of directed graphs.
• Simulate and analyze failure scenarios in water infrastructure based on the 1,000 distinct scenarios per network.
• Develop forecasting models for network state (e.g., pressure, flow) based on multivariate node and edge time series.
• Benchmark graph representation learning methods on a domain-specific, structured dataset.

Strengths

• Contains 36 distinct Water Distribution Network topologies.
• Provides 1,000 unique scenarios for each network, enabling robust scenario analysis.
• Data is structured as spatiotemporal graphs, a format directly applicable to graph machine learning.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

huggingface

Freshness

Last updated 2025-11-25 14:56:55; freshness should be verified.