Vigorously promoting the construction of audit informatization with big data as the core is the only way to improve audit ability, improve audit efficiency and audit quality, and better achieve full audit coverage. Only by deepening the knowledge and understanding of big data audit can we use new technologies and change innovative working methods to meet the needs of auditors under the background of big data development. In the big data environment, data analysis and audit projects will be more closely integrated. Audit mode is no longer a traditional audit mode, nor is it a one-man show of data, but an organic combination of the two. Good business knowledge reserve is the premise of successful data analysis, and advanced computer technology will put wings on audit work. Auditors can not only understand business, policies and regulations, but also restore business processes from data streams and find out clues to problems. Therefore, in personnel training and project arrangement, it is necessary to train more compound talents who know both audit business and computer technology, or auditors with excellent computer application level. The role of the data analysis team will be more prominent. It will analyze according to the needs of the project, change from passive to active, take data analysis as a part of the audit, run through the whole process, keep up with the mainstream of the current audit development, and promote the improvement of audit efficiency.

Second, do a good job in use planning and improve the collation of massive data.

With the continuous popularization of digital auditing means, regular data collection and reporting, unified centralized management and authorized use will be a working mode of data management in the future. It is impossible to establish a perfect data collection mechanism on a regular basis, perfect mass data collation and build a national audit data "audit cloud platform" overnight. In this construction period, which focuses on basic work such as data accumulation and system construction, on the one hand, we should strengthen data classification, improve the quantity and quality of data standardization, and build a good database; On the other hand, while ensuring data security, we should plan data access rights and make good use of the database. This requires planning from the perspective of long-term development, both long-term planning and short-term planning closer to reality, managing and using live data well and classifying data well. There is an objective contradiction between the convenience of data use and security. Only by doing a good job in data planning and perfecting the management system can we find an optimal balance between security and convenience and realize the digital audit mode of "overall analysis, finding doubts, decentralized verification and systematic research".

Third, using advanced technology to maximize the value of data mining.

First, select industries or audit projects with better informatization to conduct comprehensive data analysis, and find hidden problems through horizontal and vertical data association analysis. Give full play to the advantages of data audit in finding doubts and comprehensive analysis and refining, find clues through data analysis, accurately locate, implement efficiently, strategize and win thousands of miles, provide technical support and guarantee for on-site audit, realize the integration of front-line audit work and background data analysis, and continuously expand the depth and breadth of audit. The second is to dig out valuable audit clues from historical data. With the full implementation of the third phase of "Golden Audit Project", the construction of "National Audit Cloud" has been put on the agenda. One feature of cloud computing and big data is "focusing on relevance rather than causality". Therefore, we should make full use of financial, industrial and commercial, local tax, national tax and other data, and consider using OLAP data analysis technology in a planned way around specific audit projects. By building a data warehouse, we can deeply integrate and mine the existing data according to the theme, realize deep data mining by using advanced analysis models such as machine learning, reflect real problems with data, and improve data utilization value.