A dynamic risk-early-warning methodology of distribution system faults incorporating spatiotemporal imbalanced data distributions

•Risk-early-warning of distribution network faults on spatiotemporal scales.•Multi-source input features are combined for better feasibility.•HRLP variables in sparse datasets are explored to enhance precision and robustness.•Weights of components are decided via their impacts on whole reliability of systems.•Results can be used to optimize maintenance planning and overhaul scheduling.