Abstract: This study analyzes a 40,000 sample of primary and secondary school students from the extensive data from 12355 psychological hotline in Guangdong Province. Adopting one-way ANOVA and Chi-square test, the study has demonstrated the differences and platform response mechanisms of the three types of psychological help. Additionally, it has built a multi-factor Logistic regression model and a risk prediction model on the influencing factors of students’ mental health levels.. The results show that most students have good mental health, while a proportion of them have psychological disorders and high risk. There is a correlation between the variables of hotline data and mental health. It is found that individual, family and environmental factors all significantly affect students’ mental health. The risk prediction model effectively predicts the mental health level of students, and it identifies that primary school students and dropouts have a higher incidence of psychological problems. Academic pressure is the main cause of triggering problems. Furthermore, the study observes gender differences in different dimensions of students’ mental health. The study has noted a rising demand for psychological counseling with insufficient resources to meet such demand. The study has also found that hotline calls follow a periodic time distribution. Based on these findings, the study suggests establishing a provincial mental health database and an early warning mechanism to improve student mental health services. Data governance should also be improved to enhance the efficiency of hotline psychological services. More resources should be allocated to underdeveloped areas to narrow the urban-rural divide.

Key words: psychological hotline, data governance, mental health, risk prediction, primary and middle school students