The production and operation of enterprises are inseparable from the support of financial products. A large amount of valuable data has been accumulated in basic financial services such as accounts, deposits, payments, and payroll on behalf of others. However, one of these data is scattered in different banks, and the other is that even the data of the same bank is scattered in different departments, lack of effective integration, and it is difficult to exert synergy. The central bank credit data is the most typical representative of the endogenous data of financial institutions, and it is also the master of it. By integrating the credit data of different banks, the central bank's credit investigation has become a well-deserved guiding light in the credit decision-making of financial institutions. Payment data is another treasure. Banks have developed POS loans (loans based on merchants' POS flow) by taking acquiring telemarketing list data as a starting point; payment institutions have also opened up the way to export air control solutions based on payment data. Among them, the most representative is Alipay, which advocates "more collection and more loans" through the collection QR codes all over the line, and has issued small loans to more than 6 million offline small and micro enterprises (merchants). 3. Evaluation of data value Some of the above three data sources are primary data of production and operation, such as industrial chain data, which is a direct reflection of the operating conditions of enterprises; some are secondary data. such as government public affairs data and endogenous data of financial institutions (payment data is an exception, both financial The telemarketing list endogenous data of the institution is also the data of capital flow in the industrial chain, and it is also the first-hand data in essence), and it is an indirect reflection of the operation status of the enterprise. Whether direct data or indirect data can play a big role in the field of small and micro finance, the problem is that different data have great differences in the difficulty of obtaining and the scale of available data.