Highlights\ud A method to analyze spatial-temporal characteristics of district loads was developed. PCA was used to identify the buildings greatly affecting district load management. The features of electric loads of heating on a university campus were analyzed. Building type and operation mode greatly affect the load level and volatility.\ud \ud Abstract\ud Accurate grasp of district power demand is of great significance to both sizing of district power supply and its operation optimization. In this study, an index system has been established and visualized through a Geographic Information System, for revealing both temporal and spatial characteristics of district power loads caused by heating/cooling systems, including load level and fluctuation characteristics, spatial distribution of electric loads, and load coupling relationships between individual buildings and the district. Principal component analysis was applied to identify the buildings with significant impact on district load management. Using this method, the spatial-temporal characteristics of electric loads caused by heating in one university campus in China were analyzed. The results showed that building type and the operation modes had great effects on the level and volatility of the district electric load caused by heating. Buildings with high load levels and strong coupling with the peak district electric load, such as academic buildings, often had a major impact on the power demand of the district. Therefore, they were considered as key targets for energy-saving renovation and operation optimization. Buildings with large load fluctuation, such as teaching buildings, could contribute to the peak load shaving by adjusting the heating systems’ operation.