Big data technology is gaining a lot of attention and research in the present day. In this article, based on big data technology, we use Spark’s big data hybrid computing model to promote family style and family education under the premise of large-scale information processing. The minimum average distance of all clusters is calculated by computing the mean and eigenvectors of the Hopkins statistic. The cohesiveness and separation of the contour coefficients on the clusters were evaluated based on the mean values. We also examine the error-squared and criterion functions and use this method to verify the positive effect of promoting family traditions and education on the minds of young people. In this paper, we get from the comparison experiment of mining algorithms: Spark algorithm mining efficiency is 200-300 higher than MR algorithm mining efficiency, and the mining efficiency is superior. Especially when the log volume is large, the efficiency enhancement effect is as high as 96.88%, which is conducive to creating a good ideological and political education environment for young people by further improving the positive role of propagating family style and family education in the ideological and political education of young people.