Abstract: Quadrant analysis is an important method of outlier screening, which can find solutions to the problem through trend classification of two important attributes of things. Taking the industrial source quantity audit, and industrial output and coal usage audit as examples, the application process of quadrant analysis in the inventory and census stages of the Second National Pollution Source Census was analyzed. The target in the inventory stage was to identify the missing or underreported quantity, so the points distributed in the lower right quadrant were regarded as the abnormal objects. While in the census stage, the target was to report the outlier, so the points distributed in the upper left and lower right quadrants were regarded as the anomalous objects. Compared with the conventional methods such as direct comparison, expert experience judgment, ranking and proportion analysis, average and histogram method, quadrant analysis has the characteristics of strong objectivity, with fewer samples and simple operation. It is extremely suitable for the data audit of pollution source census, and can also make up for the lack of industry experts in the data audit.