When traditional companies meet big data

Big data can definitely be a popular topic in the present, shopping to large numbers, travel to large numbers, the number of visits to the hospital, to the large number of schools ..., as if any industry can be with big data on the edge, and it seems that everything can be big data. For many years of enterprise information-based traditional enterprises, big data for them are both clear and confused, they have data warehouse, data mining, Business Intelligence (BI) concept of baptism, but also do not understand the "big data" and the previous concept of how different.

According to the Baidu Encyclopedia, Big data, or the vast amount of data, refers to the volume of data involved is huge to the current mainstream software tools, within a reasonable time to capture, management, processing, and collation to help enterprises more positive purpose of business decision-making information. 4V features of Big Data: Volume (Mass), velocity (high speed), Variety (multiple), value. Obviously the above explanation is just a description of some of the characteristics of big data, and there is not too much guidance on how to develop big data and use big data for operational management.

This paper suggests that traditional enterprises should not only build data analysis platform, but also construct an ecological circle of data analysis in the process of building big Data system , so that everyone is an analyst, so that data analysis permeates every link of enterprise operation, and truly realizes the management mode of data operation and scientific decision-making. Building data analysis ecosystem can be summed up as "two markets, one platform", two markets refers to the business data market, analysis tool market, one platform is the analysis of view sharing platform.

A. Business Data market, making business data open and transparent.

Business data is the source of analysis, without data there is no big data. Enterprises after years of information construction, the general will have many sets of business system in operation, such as Office automation System (OA), financial management system, ERP system, etc., but these business systems are isolated isolation, the lack of integration of data, and the underlying database are professional design, high degree of complexity, Non-factory technicians are difficult to use and are often analyzed and used in the form of custom reports. Therefore, in the use of data on the data extraction is difficult, report demand response slow, data accuracy and so on.

The problem arises because of the complexity of traditional data structures, the fact that there is a huge gap between the data expected by the demand staff and the reports that the technician provides, when the data is not open to the business people and the report is used. Building the business data market is the data of each business system extraction, cleaning, integration, according to business process re-integration, packaged as a unified granular, unified dimension of the database table . These datasheets are characterized by the ability of business people to understand and interpret the analysis so that business people can find problems in the data in the early stages. In the process of building the data market, there are several points needing special attention:

L Adopt dimension modeling method, with "easy to understand, ease of use" as the principle . Data in the business data market is for enterprise analysts to use, they are more than technical business personnel, data structure should be consistent with the majority of the intuition, rather than the traditional "three paradigms", need to rely on complex ER diagram to express clearly. For the professional IT staff, especially to resist the design of the dimensional model as a "snowflake model" to save data duplication, increase the temptation of flexibility, resulting in the cost of model complexity;

L fully comb the business to realize the interconnection of data . Traditional enterprise business system, most of the chimney-type software system, if not at the enterprise level of business carding, it can not be very good to integrate data, extract valuable information. In the process of integrating data, we must have a unified dimension to effectively achieve data integration, such as unified customer code, ID number, etc., so combing the unified dimension of enterprise level is the key step to realize data integration;

L Sub-theme, steady progress . There are many business processes involved in the operation of the enterprise, cram is impossible. In the construction of enterprise data market, according to the characteristics of the enterprise, priority construction of the operation of the most important business, as soon as possible to put into use, gradually improve, for example, a sales class enterprises can first integrate the order business;

two. Analyze the tool market to create an analytical framework that fits your needs.

The data is a gold mine, but there is a need to rely on good tools to extract the gold. Direct analysis of raw data although flexible, but more suitable for high-level personnel, for most of the frontline staff, should provide more easy-to-use analysis tools, such as reports, charts, reports, etc., the business indicators in an image of the way out. The analysis tool market is the place where the analysis tools are pooled within the enterprise and are provided to the employees according to their own circumstances.

The analysis tool market consists of two parts, one is the development platform for analyzing tools, the other is the portal platform to run the analysis tools. In the process of building the analytical tools market, there are several points to note:

L Analysis Tool market construction . From a technical point of view, the analysis tool market can find a large number of software systems to meet the needs of the enterprise can be based on the actual situation, or procurement, or use of open source, to build such a set of software systems, not only to enable enterprises to customize their own reports, analysis charts, but also to make good analysis of the report show out;

L Analysis Tool development to the enterprise itself, outsourcing supplemented . Analysis tools and business characteristics, and people use habits and other factors closely related to the requirements of high frequency of change, all rely on external manufacturers to develop, on the one hand, the cost is high, on the other hand, the demand response is not timely, it may be developed, the analysis of the demand has not, so enterprises should cultivate their own development team, Can use data in the data market to develop a variety of query reports, analysis charts, etc.;

three. Analyze perspective sharing platform to make data analysis social.

In the age of big data, there is a scarcity of valuable analytical results at the same time that data is plentiful. Data analysis, is definitely not high above the spring snow, but the need for everyone to participate, the atmosphere of contention, to establish in the enterprise everyone is an analyst concept.

The analysis of the building of view sharing platform, the use of social media such as Friends Circle, Weibo and other mechanisms, so that everyone has the opportunity to express the analysis of the data views, through the forwarding, commenting mechanism, so that a valuable point of view to float up, in the process, sharing views of individuals can get a sense of accomplishment, forwarding, commenting people have a sense The enterprise obtains the idea which has the value to the operation decision, and realizes the scientific and democratic decision-making based on the data analysis process. In the process of building a sharing platform, there are several points to note:

L Personnel attention mechanism . Because the enterprise operation is different from the Internet, there are certain closed characteristics, not recommended the use of Weibo attention to openness, but should use a friend circle similar "concern-consent" of the friend mechanism, to avoid inappropriate attention to bring about the disclosure of information;

L share range control . Because of the particularity of enterprise data, employees need to control the scope of sharing and avoid the leakage of confidential information when analyzing viewpoints. Control the scope of sharing, both manual control, and from the platform level through the authority control, such as financial data sharing limited to a certain number of departments or personnel;

L share the idea sorting algorithm. combine various factors to sort out the published ideas so that what is present in front of everyone is the most desirable content, creating a good environment for sharing, and avoiding the devastating impact of high-quality ideas on the ecological circle.

Believe that enterprises through the construction of "two markets, a platform", will be able to fully discover the value of data, in the tide of big data, to truly achieve data operations, scientific decision-making management mode, to avoid the mistakes of experience decision-making.

When traditional companies meet big data