Design flow for BI Data systems

BI Big Data Products:
The data management platform can be built through a report or BI module. In the column "Sail Soft data Application Research institute" there are examples of enterprise data management and BI reporting platform construction.

From the point of view of project implementation, it can be considered from two aspects of technology and business.

Prior to the demand research. A map of building ideas is listed.

Technology needs to consider two major subjects, one is the company's data volume, but the specific demand situation.

Data conditions

1. Data source: What are the business systems that consider the main sources of data? Whether the data warehouse needs to be established. and pre-preparation work, such as related data dictionary preparation, and even the number of interfaces to take the problem.

For example, financial data from EAS, management data from OA, sales data from SAP, and database types for each business system.

2, data situation: The largest single-table data volume and data increment in the business system. By understanding the maximum single table data volume and data increment, the preliminary Planning data processing scheme, whether to establish the middle tier, index, GP and so on, choose the appropriate analysis platform, is BI or general reporting platform.

For example, the sales data sub-table already has 2000W data, the monthly increment 50W, then the processing plan may consider through the report platform Finereport to display, establishes the middle tier and the index, enhances the query efficiency and the display performance. or through the Big Data bi platform Finebi set up the foreground analysis, adopt Fineindex way, to improve the display performance.

3, data repair: The main dimension of data maintenance and factual data to record, the specific work is to fill out the maintenance of the page. For example, labor analysis in the wage analysis, usually financial offline data, then need to develop the import page for data entry.

This work will need to consider where to record the data recorded: If the Data Warehouse is built directly into a number of warehouses, but if the development of the report through direct connection, it is necessary to record the recorded data directly into the business system.

4. Data Warehouse: To determine whether the data warehouse needs to be established according to the demand subject, analysis index and data condition. If you need to integrate financial data, expense data analysis, but the financial data from EAS, the cost data from OA, then the need to establish a data warehouse for integration analysis. It is necessary to consider whether the corresponding relationship between the relevant dimension data and the staff organization structure can be effectively mapped after the Data Warehouse is established.

Demand situation

The main requirements analysis is the identification of business indicators, subject modules and analysis scenarios.

1, the early need to determine the platform for the application of the target, different levels of leadership attention to different indicators, need

Plan analysis index and analysis form. For example, the leader only focus on sales, profit margin and other key data, and the business layer on the basis of indicators may also involve drilling analysis, linkage analysis.

2, the main body of the demand consists of three parts: Demand module + module indicator + analysis form. Through the requirements Module + module indicator + analysis form, initially framed the requirements, assessment of workload and implementation of the quotation. For example, sales analysis, involving customer analysis, order analysis, product analysis, channel analysis and other analytical indicators, each index may require a level drill analysis, analysis and other forms.

Difficult point

First of all, for such a platform to have a good positioning-business-oriented agile corresponding development platform.

Secondly, how to deploy the personnel, how to promote the smooth progress of the project. Personal advice is that if people can fully try to get frontline business people to participate in or even directly engage in system development, it can improve development efficiency, meet business requirements, and reduce it pressure. Second, must have the leadership of the high-level personally involved, from the top to promote implementation.