Data Warehouse - Mining 01-5-21 in the "Beer and Diaper" 04:19:25
interlocutor:
Moderator: Cheng Hung - "Computer World" reporter
国辉 - "Data Warehouse Road" website host
Dialogue topic:
What is the application prospect of data warehouse in China?
How high is the threshold of the data warehouse?
Moderator: Since the story of "Beer and Diaper" is widely circulated in China, the data warehouse is also lively in China. Many entrepreneurs have a great hope to it, but why short-term data warehouse applications enter "Hemakerge"? What factors have been made in China?
Meng Xiaofeng: Data warehouse is developed based on the database, which usually has three parts: data warehousing, online analysis processing (OLAP), data mining, which has strong complementary relationships. Data warehouse is to meet the needs of data analysis on the accumulation of data accumulation on high data accumulation, but due to the shortcomings of China's accumulation of basic data, the application of data warehouse technology has not been promoted. The emergence of a technology is generally due to new concepts, researchers solve certain problems; although the data warehouse has across this phase, but currently in the application promotion process of the data warehouse, I have encountered a threshold. That is how this technology is accepted by most people. I think that the data warehouse is accepted by most people, and the problem that needs to be solved is to combine data mining technology with existing commercial technology. At present, there are many applications in the market to solve the universal data mining system for all business models, but in fact these systems are extremely difficult, only those talents who are very familiar with data mining technology can understand and use them, and ordinary users are difficult to apply These techniques solve their own business problems.
Yang Shun Sheng: We are a company that truly operates data warehouse products. From our experience in contact with domestic enterprises, data warehouse technology has failed to develop in China, mainly with the following reasons: First, China's informationization The basic equipment is relatively unpleasant, for example, the current charging data of the telecommunications industry is very dispersed. There are 40 ~ 50 species of the billing tool, and the data collection is all kinds, bringing many technical difficulties for future analysis. Second, the competition awareness and service awareness of the company are not strong enough. It is not so urgent for decision-making analysis. Since the company does not have the opportunity to implement data warehouse, it is also possible to cause the lack of technical personnel in this regard. Third, the data warehouse is a system shared system. The information obtained from different levels will be greatly different. It is a good tool for business decisions, but the current Chinese company has not established a management mechanism to promote data. Sharing, whether it is the ability to people, the organization of the company is not a continuous management mechanism, whether it is very difficult to establish data analysis on this basis.
Ji Zhihui: In fact, in an environment in which the foreign market competition is very intense, each shopping mall has already taken a way, many of the laws that have been artificially discovered have long found, in which case data mining Technical to solve the problem is a very normal idea. However, data mining tools don't directly tell decision makers to put beer and diapers together. There are many domestic enterprises to implement data warehouses, but most of them are not satisfactory, their key reason is that when building data warehouses, be confirmed to invest in data warehouse? What is the problem of using data warehouse? What is the data warehouse? The purpose can you tell a paragraph? Otherwise, don't know how to use it, how to evaluate its success or not. Second, the data warehouse is not the software product that can be used. In fact, the data warehouse is more like a process, and a user will gradually know yourself and improve your own process. Third, the data warehouse not only reflects the current state of the company, but also rely on the user to make the final decision. In summary, the reason for the less successful domestic data warehouse is in addition to the user's application level, the business management level needs to be improved, and the price of data warehouse products is high, and there is less affected to its successful promotion in China. Chen Daobin: I am working in information analysis, and some research is also launched for why the data warehouse is also launched. As a user, I feel that the banking industry needs data warehouse technology, and it should most develop data warehouse technology. In recent years, several domestic banks have tried in this regard, but there are many failures so far, and there are fewer success. The most important reason is that many banks are not clear when establishing a data warehouse system. A data warehouse system should distinguish between business processing systems, because the requirements of business systems are often simple response and interface simple. The data warehouse and the business system are not parallel, it should be based on all business systems, to collect, analyze, organize, and publish the business information, etc., should be a stable, data collection with time parameters. Data warehouse technology itself has no new content, which is a large fusion of management science, computer science, network science and analysis.
Is data warehouse technology easy to use?
Moderator: Is there any technical reason for data warehouse applications? Is there contradiction between the user's level and the front end tool of the data warehouse?
Meng Xiaofeng: In the three concepts of the data warehouse, the data warehouse is the basis for the data analysis of the company. Its main job is to summarize the raw data in the database, and aggregate into a data set that can be used for high levels. There are two types of analysis tools on the data warehouse, and one is OLAP for analytical work, and the other is data mining of predictive work. The proposal of data mining concept is to find a law that is associated with "beer and diaper". But the current technical system, whether in China or other countries in the world, the application is subject to certain constraints, the most important reason is that this technology has not reached the maturity and ease of use of database technology. But in the current view, the availability of all products is also worthy of doubt, because if you are not a database expert, statistical experts and artificial intelligence experts, you will be difficult to use such an analytical tool. At present, data warehouse products on the market are designed based on a common technical platform. Although this product can solve different users' analysis needs, it does not integrate commercial logic of special fields with data warehouse technology, and thus analyze the effect It is impossible to achieve peak. Another technical bottleneck is currently an endless way, and has not experienced a precipitate of a large wave, and the search technique in the database has been a few years of exploration. It has formed several fixed, mature technical models, which is not possible to achieve data warehouse products. Another reason for the practicality of database product, so the development of current data warehouse technology is still in the accumulation phase.
Chen Daobin: Industrial and Commercial Bank's unified understanding of the establishment of a data warehouse system is that the data warehouse products that can be purchased directly in the market must be customized according to their own business, they must clear their data sources and business needs. Then, then the work of the middle bridge is good, this bridge requires the support of the data warehouse product in the market, and must pay attention to the coordination of business and technology from the beginning. Ji Guohui: Since the data warehouse is generated from Western countries, it is a strong Western cultural color, the most typical is the report in the data warehouse. Foreign products pay attention to the contents of the report, but China requires the same content and format, and even more important than content, at this point, foreign report tools are difficult to meet the needs of Chinese users. As a tool, each level of the company can use it to generate benefits, but truly realize, or there is a user level and demand problem, we can't ask each user to get data from the data warehouse, And security measures are not allowed to do so. At this time, we need a series of different data warehouse front-end tools, which is exactly the lack of all data warehouse products. Most of the current products only provide a tool, trying to meet the needs of everyone, and everyone is not satisfied.
What is the application range of data warehouse?
Moderator: Which industries have a big demand for data warehouses? What industry applications are currently available in China? why?
Yang Shun Sheng: I have analyzed the maturity of certain industries and enterprises in the implementation of data warehouses in some industries in accordance with some hypothetical conditions. In the 500th list of global wealth magazines, nearly 50% of companies have implemented enterprise-class data warehouses or departmental data, we learned from telecommunications, bank, retail, air, railways, postal, food, consumption Manufacturing, automotive, medical, insurance and other industries are the most important industries for data warehouse technology. In all of these industries that implement the data warehouse, the proportions are: 17 retailers, 16 aviation industry, 15 cable phones, 14 mobile communications companies, 13 banking industry. In addition, we also have statistics on the number of enterprises currently implementing data warehouses in different industries in different industries: According to foreign implementation, we found that leadership needs, information technology infrastructure, analytical application needs, and intensive levels of competition The four factors are key factors affecting the progress of the enterprise to implement data warehouse. The larger the company, the more historical data, the higher the urgency of the implementation of the data warehouse; the retail and manufacturing enterprises currently pay more attention to cost control, will first use operational and production analysis, the government supervision department is collected due to historical data. Difficult, implementation of the data warehouse is relatively slow; enterprise group headquarters to strengthen supervision, will adopt data warehouse as information technology means for processing and analyze the concentration; leadership of business management education, pay more attention to scientific decision-making means, therefore Compare the construction of data warehouses.
Chen Daobin: At present, only the Chinese financial system has only breakthroughs in the application of the ICBC in the data warehouse. The reason is: First of all, the business banks are large, the customer group is huge, so they must conduct in-depth research on customers to achieve customer-centric Service mode; Second, As early as September 1999, the ICBC proposed to handle all the business in Beijing and Shanghai two centers. In fact, the data concentration necessary to establish data warehouse; third, ICBC has a biggest advantage that has implemented a unified development of integrated business system, providing customer information integration conditions. From the perspective of leadership support, the current sectors are specially studied commercial development problems in high-tech environments. He has unique insights with the use of information technology development banking, so ICBC currently in the data warehouse in the first project. (Customer Relationship Management) has a substantial progress.
How to cross the data warehouse threshold? Moderator: Is China Data Warehouse Market mature? What can we make this solution for the previously mentioned constraints? What kinds of measures should I take to promote the application of China's business data warehouse?
国辉: This question is good! The Gartner Group has had a report of data warehouse market, which can be seen from the report that by 2003, the US data warehouse sales will account for 58% of the world, and Asia accounts for only 7.5%, it is not difficult to see our gap. . But the current technology development speed is similar, so it is said that the data warehouse application market is still present in China. I think the only solution is to let the enterprise directly facing the fierce market, changing management ideas, so soon, it will produce demand, reading books and boxing will always understand the rich experience of the boxing.
Meng Xiaofeng: Now more and more companies are building web-based electronic stores that can collect a large number of raw data, so e-commerce has become a very promising application area of data warehouse technology. How to provide a special data warehouse solution for e-commerce applications, which should be accepted by a common solution, thus crossing the threshold for the application process of the data warehouse. Now there are many customized needs in the field of data warehouse applications, and users are urgently needed to provide data analysis and tools that provide customer personalization analysis.
Yang Shun Sheng: From the experience of advanced countries, we found that online transaction processing system (OLTP) and enterprise networks two information infrastructure is a must-have condition for implementing data warehouses. The more intense the competitive environment, the more data warehouse system, the company needs to understand the needs of customers, and it is necessary to find business risks while still need business analysis and management. All of these analyzes are the analysis of large data, and traditional information technology will be produced. Many limitations must rely on a TB-level data warehouse system to solve the above problems. Data warehouses are analytical applications and is the most suitable information technology that solves complex business problems. But do these perspectives are suitable for China's corporate ecological environment? Is there a domestic case to support the above view? Recently, Shanghai Securities Central Registration Co., Ltd. and China Civil Aviation Information Network have successfully implemented TB-level data warehouse systems. Both cases are the most powerful arguments for the need to implement data warehouse in China and institutions.
Chen Daobin: From the situation of implementing data warehouses from the Industrial and Commercial Bank, data warehouse technology has great development prospects in China. The technical and method system of the data warehouse itself has a lot of technology, but when establishing a data warehouse application, it should also be guided by the problem-oriented without the way, and the products and tools are found according to the problem. There are too many examples in this area. The main reason is that many companies are purchased by the data warehouse product. I feel that this product is very good. I must use it, and then I have begun to build my own system. This practice has been practiced. Proof is not feasible. Developing data warehouse applications cannot be eager to seek, I hope that all problems can be solved once. The construction process of the data warehouse should pay attention to methodology. It is necessary to do in a large demand framework, and solve a problem in the case of business and technicians.
[Reporter Review] "Gunner" and "Gun"
During the dialogue, the biggest feeling of the reporter is that because the data warehouse does not directly tell the decision maker, the beer and diaper should be put together, so the company cannot press all "treasures" on the data warehouse. If the data warehouse is a good "gun", the decision maker should be "playing people".