Three tools needed by the hottest Bi landing

Three tools needed for the implementation of business intelligence (BI)

users in various roles of business intelligence (hereinafter referred to as Bi) must rely on and use tools to meet their needs. Bi roles are divided into three categories: business, business integration technology and technology. For their corresponding application requirements (including deductive and inductive) and control development requirements, deductive requirements support tools, inductive requirements support tools, and control development tools are needed to realize them. This article first describes seven tools of deductive demand support, which can be divided into descriptive statistical tools, business technology and methods, economic forecasting methods and models, OLAP analysis, knowledge discovery tools, expert systems, and decision-making methods and models. Control development support tools generally include system management tools and development tools

the author believes that Bi takes epistemology and organization theory as the basic principles and adopts appropriate "tools" to help "relevant roles" make the best decisions on "relevant contents" within their scope of responsibility. It is composed of "three-dimensional model" and "three-tier funnel". It is an overall decision support scheme that assists the whole enterprise to integrate concept, organization, process and technology. The three-dimensional model is composed of role dimension, content dimension and tool dimension, reflecting the general principles of Bi, such as subject, object and tool. According to the definition of Bi, only by clearly dividing the relevant roles and determining the needs, and with the help of tools, can the purpose of Bi auxiliary subjects to make the best decisions on the matters related to the scope of responsibility be achieved

bi's business role, technology role and business integration technology role correspond to business application requirements (deductive requirements of business type and inductive requirements of business integration technology type) and technology application requirements, that is, management and control development requirements. Whether this demand can be effectively realized and promoted must rely on tools. For the two needs of the three categories of roles, Bi tools are divided into application-based support tools and management and control development support tools. In view of the scope of the article, please refer to relevant books for the content of hardware support tools

application support tools can be divided into descriptive statistical tools, business technology and methods, economic forecasting methods and models, OLAP analysis, knowledge discovery tools, expert systems, and decision-making methods and models. Control development support tools generally include system management tools and development tools. Generally speaking, these tools form products in the form of software packages. In view of the important role of Bi for business applications and the significance of successful business applications to Bi, this paper focuses on the analysis of Bi application-based support tools, and makes a brief introduction to relevant products

first, there are seven tools to realize deductive requirements. The deductive content of Bi can be divided into three levels: report query, comprehensive analysis, decision-making and selection. As shown in the figure below, the deductive needs of Bi are realized through descriptive statistical tools, reports and display tools, economic forecasting methods and models, business technology and tools, OLAP analysis and expert system tools, decision-making methods and models

in the above figure, descriptive statistical tools help users fully understand the facts at the report query level; Comprehensive analysis helps relevant subjects to seek reasons or obtain suggestions directly for simple problems in a logical way, which needs to be realized by using economic prediction methods and models, business technologies and tools, and OLAP analysis. Expert system and decision-making method and model are powerful tools to realize quantitative and partial qualitative decision-making. Through this kind of tool, the user can help the subject to choose the best by evaluating the advantages and disadvantages of each scheme when making decisions and choices, and draw a conclusion. The three-level structure of Bi deductive application reasonably and step by step solves the transfer and application of the deductive thinking method of premise, logical rule and conclusion

(I) descriptive statistical tool

the basic significance of statistics is to use statistical indicators to study the quantity and quantitative relationship of socio-economic phenomena through the comparison and development of index values, and to show the process, results and laws of their changes and development. As one of the two basic contents of statistics, descriptive statistics is the basis of the whole statistics and the first step of statistical research. It includes the collection, collation and display of data, the extraction and analysis of useful information in the data, and the analysis of the centralized trend and decentralized trend of variables is its main content. Descriptive statistical tools refer to the use of basic indicators to meet the needs of relevant subjects to understand the basic facts. These basic statistical indicators include total indicators, relative indicators, average indicators and variation indicators

total index: the value indicating the overall scale, level or total work volume of specific socio-economic phenomena, which is the basis for calculating various derived indexes. For example, the annual sales volume of a group company is the total index. It can be divided into total unit volume, total mark volume, total period volume (such as the total sales volume of a business department), and total time point volume (such as month end inventory). The total index can be calculated directly and indirectly

relative index: a statistical index indicating the quantitative comparative relationship between two interrelated social and economic phenomena. Such as the comparison of inventory turnover between competitive enterprises, or the comparison between the sales volume of an enterprise in the industry and the total sales volume of the industry, etc; The commonly used year-on-year and month on month comparisons are also the use of relative indicators. Relative indicators reflect the essence of social and economic phenomena and their quantitative comparison relationship, and clearly understand things from the quantitative comparison relationship between phenomena. Secondly, we can find a common basis for scientific comparative analysis of phenomena that cannot be compared directly. Relative indicators include: plan completion index, structure relative index (the respective proportion of the profits of high, medium and low-grade products in the total profits), proportion relative index (the comparison between product a and product b), comparative relative index (the comparison between the sales growth rate of an enterprise and the sales growth rate of competitive enterprises), intensity relative index (such as the comparison between the article/person index in the sales situation of tobacco industry) Dynamic relative indicators (such as the comparison of sales over the years). When using, we should correctly select the comparison base, determine comparable comparison indicators, and use relative indicators in combination with the comparison base

average index: it reflects the general level of a certain quantity mark of each unit in the homogeneous population, and can be compared with the general level. For example, the comparison between the annual average sales volume of the branch company and the average sales volume of the head office can be used for quantitative calculation and prediction. There are five kinds of average indicators: arithmetic mean, harmonic mean (less used, mainly used when the total amount of signs is known but the total amount of units is not known), geometric mean (generally used to calculate the average development speed), mode (the number with the most times in a group of series), median (the number in the middle of the data arranged in order of size). When using the average, it should be combined with the variation index, and sometimes it is necessary to supplement the overall average with the group average

marker variation index: it indicates the difference degree of a certain marker value in each unit of the whole. The variation index can reveal the difference covered by the average index, and can also be used as a measure of the representativeness of the average index. It can be divided into: full range, mean difference, standard deviation and coefficient of variation of signs. The full range is the difference between the maximum and minimum of the mark value, such as the difference between the highest and lowest sales of a product in all provinces of the country in 2006; The mean difference is the absolute value of the deviation between each flag value and the overall arithmetic mean. The standard deviation is the positive square root of the sum of the squares of the deviations between each marker value and the overall arithmetic mean. The larger its value is, the greater the difference is, and the average value cannot well represent the general level; The ratio of the standard deviation to the arithmetic mean is called the coefficient of marked variation, which is used to compare the two mean indicators when they are not equal

descriptive statistical tools are very common in the current industry application of Bi. It is worth noting that the current report query application of general Bi involves more total indicators, average indicators and comparative indicators of descriptive statistics, while the indicator variation indicators that describe the deviation trend are not used enough. In fact, the indicator of logo variation can solve many substantive problems of Bi's application subjects, such as finding brands with unstable sales among many brands and customers with large trading volume fluctuations among many customers. These information plays a key role in the research and development, production, procurement, sales and inventory of Bi's application subjects

(II) report and display tools

are distinguished by report application categories. BI system can use report and display tools to generate statistical reports and query reports. The query report is relatively simple and can be easily customized according to the needs of users. Due to the characteristics of many header items and one or more slashes in the header of Chinese statistical reports, and even the complexity of cutting sub headers in the header items, it is difficult to generate statistical reports. Compared with some domestic products, such as crystal report, the operation process of foreign BI products is more complicated. For the display of reports, one is to generate tables, and the other is to generate graphics corresponding to tables, such as curves, column charts, three-dimensional charts, etc. Generally speaking, reports are required to be linked with graphics. That is, when the report data changes, the graph changes. Or on the contrary, changes in graphics will bring changes in report data. This is easy to happen during drilling. For example, the business intelligence of SAS and the graphic interaction function of Bo can meet this requirement

report and display tools focus on realizing the flexibility of reports, emphasizing the beauty of graphic display, the templating of graphic color, and the diversity of graphic types. This is the basic function of Bi tools. When we talked about Bi in the past two years, it was easier to judge whether the interface of a manufacturer's bi was beautiful, etc., which was caused by people's insufficient understanding of Bi applications at that time. Now when promoting Bi, manufacturers not only pay attention to the deeper comprehensive analysis function and prediction and decision-making function, but also expand to the full set of Bi platforms such as data integration and integration, data storage and metadata management

(III) economic prediction methods and models

economic prediction methods and models are the application of statistical inference in statistics, which are complex and flexible. It is the application of Bi at the level of comprehensive analysis. This method meets the needs of business users looking forward to the future. Statistical inference generally includes parameter estimation, hypothesis testing, classification and selection. Economic forecasting method is the application and promotion of parameter estimation, including point estimation and parameter estimation, such as the predicted value of sales volume in a certain period of time, or the interval in which the predicted value is located. The economic prediction model is a modified prediction model based on the actual situation of the industry under the guidance of the prediction method

quantitative prediction methods many models can be established in practice. Here, time series prediction is mainly used to briefly introduce the application of economic prediction methods and models

time series prediction method is an application of dynamic analysis method. Dynamic analysis is a method that compares the number of economic phenomena at different times in statistical research, so as to understand the direction, speed, trend and law of phenomenon changes, and predict the future based on it. Dynamic trend analysis and prediction is an important application of dynamic string 5 analysis. Time series is a major aspect of dynamic analysis and research. Its premise is to compile time series and form time series prediction method. There are about 100 prediction methods in the current time series, but its basic density is 250 ~ 30