In the era of deep integration between mobile internet and the digital economy, users' "digital footprints" — unstructured data encompassing online transaction preferences, social interactions, and other dimensions — are evolving into high-value information assets in financial risk management. Traditional credit evaluation systems, which primarily rely on static structured financial data, face inherent challenges such as information asymmetry and limited population coverage. This study examines how Ant Financial, a leading fintech enterprise, leverages massive user digital footprint data to construct innovative risk management models. The research develops a simulated dataset containing 50,000 samples that integrates traditional financial characteristics with digital footprint indicators reflecting user behavior. Two models are then designed for comparative analysis: a linear regression model based solely on traditional financial features serves as the benchmark, while a decision tree model incorporating digital footprints is implemented through feature correlation screening. Through comparative analysis of these models, this study demonstrates the pivotal role of digital footprints in expanding the reach of inclusive financial services and enhancing credit risk prediction accuracy.Research demonstrates that decision tree models incorporating digital footprints outperform traditional approaches in identifying potential default risks. The model structure also reveals soft information such as user behavior stability and consumption patterns, which holds significant value for credit assessment. This provides valuable insights for the digital transformation of traditional financial institutions. Furthermore, the study offers an analytical framework to understand the core risk control logic of fintech companies like Ant Financial.
Yu Xinbo (Wed,) studied this question.