Water Quality and Rainfall Time Series for Anomaly Detection

High-frequency water quality, water level, and rainfall data from three monitoring stations in China's South-to-North Water Diversion Project. The dataset supports an integrated framework for water quality monitoring and abnormal early warning, developed by Xu Chen. It was last updated in April 2026.

Use Cases

• Train a Stacking ensemble model to predict future water_quality values from temporal patterns in water_level and rainfall.
• Apply rolling-statistics-based detection to water_quality sequences to identify and exclude isolated false data points.
• Use rolling-variance screening on variance-shrinkage sequences to detect sustained structural deviations.
• Implement difference-threshold rules on rapid-change segments in the time series to classify true anomalies.
• Validate knowledge-data coupling mechanisms for reducing false alarms in operational monitoring systems.

Strengths

• Data originates from three distinct monitoring stations, providing spatial context.
• Includes multiple correlated features: water quality, water level, and rainfall.

Limitations

• The dataset is very small at 5.5 KB, indicating limited sample size for model training.
• Specific row counts, column names, and measurement frequencies are not provided.
• Geographic coverage is limited to the Eastern Route of China's South-to-North Water Diversion Project.

Provenance

Source

Monitoring stations in Phase II of the Eastern Route of China's South-to-North Water Diversion Project.

Collection Method

High-frequency monitoring data collected from sensors.

Time Range

null

Freshness

Last updated in April 2026.

Geography

China's South-to-North Water Diversion Project.