(a) Diversification of data sources. With the advent of the big data era, Internet credit data is no longer limited to bank credit data. On the one hand, the core business of Internet companies provides a wealth of Internet credit data, such as Sesame Credit, which relies on credit data from Ali's e-commerce transaction data and customer evaluations, and Tencent Credit, which relies on social network data from QQ users and WeChat users of social platforms. These data include social, political, commercial, cultural and health information, which can be used as weak variables in credit assessment models to analyze the essential behavior of customers and predict their future trends. On the other hand, P2P companies build their own credit database. With the development of P2P, P2P enterprises have collected more and more credit data, such as PaiPaiLi, Renren Loan, Xinzhifu and other enterprises have established their own credit system.

(ii) Extensive coverage. At present, China's credit system covers about 800 million people, but only about 37% of them have credit records, and the other 500 million or so do not have bank credit records. For the financial sector, these 500 million people are potential customers of banks, and a favorable opportunity to expand business. How to understand the credit situation of these 500 million people has become a challenge for all financial institutions. However, with the development of the Internet, more and more individuals and enterprises have left a lot of information on the Internet, broadening the scope and source of credit data, and the application of big data and cloud computing facilitates the collection of credit data. The use of collected data to make credit judgments on those who have no credit records can meet the needs of financial institutions for pre-credit review of borrowers' willingness and ability to repay.

(3) Data collection is hidden. Compared with traditional credit collection, Internet credit collection enterprises in the era of big data have their own rules for data collection, collation, processing, analysis and use. First, the processing of collected data by Internet enterprises is usually carried out by means of cloud computing and background encryption technology, and the information subject is not clear about the scope of collection of its own network information due to technical reasons and information asymmetry. The second is that many Internet companies currently stipulate in their privacy policies that they have the right to exchange and cooperate user information with other platforms that abide by the same privacy policies, and users are prone to ignore such information.

(iv) Enrichment of application fields. First, the content of credit data is more comprehensive. For example, Sesame Credit collects information from five dimensions, namely, customers' credit history, behavioral preferences, fulfillment ability, identity traits, and personal connections, which provides an important basis for financial institutions to evaluate the credit of loan recipients and decide whether to issue credit. Secondly, the application of credit evaluation results tends to be more life-oriented and daily. In addition to borrowing and lending, the application field has been expanded to a wider range of uses such as accommodation and traveling. Third, it provides an effective basis for decision-making by regulators. After mastering these big data, the regulator can understand the situation of credit changes of economic subjects in a timely manner, make more sensitive camera choices, and monitor the flow of funds to check the implementation of policies to achieve the accuracy of regulatory policies.