Modeling and Analysis of Group User Portrait through WeChat Mini Program

Abstract

Meeting users’ preferences and increasing business revenue is an ongoing challenge in the mobile service application. In this paper, we address these challenges by mining mobile user behavior patterns and propose an approach to construct a group user portrait by analyzing access data collected from the users of the WeChat Mini Program. We extract the attributes of mobile users considering their geographic information, online duration, and age group. Using$Z$-score standardized processing and$K$-means clustering algorithm, we then model the user portraits through three dimensions including daily average duration, interaction intensity, and access frequency. Our analysis has two important features. Firstly, the significant log data used in our experiments was collected from the production environment ensuring that the results reflect the real attributes of WeChat Mini Program users’ behavior. Secondly, we provide data-driven decision-making to help marketers enhance the quality of the product and improve user experience. The experimental results indicate that by distilling and analyzing the key factors from the log data, the characters of typical users can be properly profiled to help product owners better optimize the exact set of the features which need to sustain and further grow.