Analyzing public service survey data and developing quality improvement measures is a common measure used by government departments to improve public satisfaction. Due to the influence of economic development, urban construction and other factors, the attention of different people to the area of public services is different. Simply focusing on the value of satisfaction may lead to a mismatch between the strategy of the government and the public's concern. In this paper, a neural network-based analysis approach is proposed, which is capable of mining the focuses of the interviewed public from the satisfaction survey data. In this approach, the neural network technique is used to train the relationship model of service areas and satisfaction indicators, and then the SHAP technique is employed to analyze the ability of each service area to influence the satisfaction indicators. Based on the proposed approach, this paper analyzes the 2023 public service survey data of City A in Hubei, and finds that service areas such as public sports are more popular among the local public.