Test run analysis Report of a company based on Finebi data decision platform

First, the software and hardware environment of data platform

650) this.width=650; "src=" http://img.blog.csdn.net/20160129144600417 "alt=" here write a picture describing "title=" "style=" border:none; "/ >

II. Organizational structure and authority system

A set of organizational structures can be established in the organizational platform to create departments and personnel. You can also import your organization into the platform by integrating with your existing system's organizational structure.

Feature permissions enable each user to have corresponding BI access rights by configuring the function point URL. The first time a user accesses a feature point of a protected resource, an access-request server is issued with a request that verifies the user's permissions and returns the login page if it is not authenticated.

Data permissions Many systems have permission authentication only to restrict the use of the module so that legitimate users can exercise their rights. In order to satisfy this kind of overall authority authentication, the control force can achieve the same BI content in different permissions to show the effect is not the same as this is to avoid the production of a large number of BI to achieve the same effect, especially in the enterprise internal business complex approval of the trouble when a bi can solve all problems. The platform controls the permissions of different users on the data from the packet level to achieve the fine-grained control of the data through the permission control of the business package.

Level of permission control

1. The data permissions of different users for the data business package are limited to the scope of their own permissions.

2. Different users can access reports within the scope of the permission.

3. Different users only have access to the data within the permission range for the same report.

Third, data processing

The data source supports data sources such as Oracle,db2,sqlserver,mysql,sqlserver,informix. Support for ODBC data sources support for shared application server data sources. Supports the program data interface. Supports text data sources. Supports built-in datasets.

A business packet, or cube, is the basis of data for immediate analysis. The data business package is created by the Data Administrator, which contains all the business data tables, datasets, interface data, text data, etc. that can be provided to the analyst. stored in the server directory in the form of a file (suffix named Fcube).

650) this.width=650; "src=" http://img.blog.csdn.net/20160129144611652 "alt=" here write a picture describing "title=" "style=" border:none; "/ >

Data Escaping
You can escape the table names and field names in the data business package as data escapes that are understandable to business people can be directly integrated with annotations in the database or can be edited directly manually.

Data Association

The association between data is used to establish inter-table relationships between multiple tables or to directly inherit foreign key relationships defined in the database. Tables that participate in association relationships must have primary key support.

650) this.width=650; "src=" http://img.blog.csdn.net/20160129144620809 "alt=" here write a picture describing "title=" "style=" border:none; "/ >

Iv. Data Conversion

Data Conversion provides a variety of tabular and multi-charting services to showcase data with a variety of business needs. List, group, and crosstab chart types include column charts, bar charts, pie charts, area charts, combo charts, dashboards, and maps.

new columns enable data conversion and data calculation through custom data columns. Includes building self-looping columns based on existing data new columns build custom data columns based on formulas. Easier to use for later analysis.

new columns are created based on existing data columns based on the existing data column and are formed by custom grouping. Primarily used to establish common custom groupings for use by all analysts.

building a self-looping column provides a hierarchical hierarchy of organizational structures based on a single column of organization IDs in a single column of databases, or a hierarchy of organizational IDs and parent ID data in a two-column database. Mainly used for organizing tree display.

New formula column formula engine supports data type conversions common functions, math and trigonometry, text functions, date and time functions, logical functions, array functions, report functions, and other custom functions
。
row and column conversions are primarily used to combine the field values of one of the columns in a database with other metric fields into a new field.

Five, the regular update

The data platform automatically generates the corresponding cube in the background when all the data is set up in the full volume update . The cube of the data platform has the form of MOLAP, so it has excellent support in dealing with the problem of large data volume. The data in the cube can be set to a timed full-scale update.

incremental Updates can only be set separately for tables and datasets in a business package. And only valid for new data.

VI. Data analysis

Dimension Metrics Analysis provides the flexibility to select any metric and dimension from the business package for autonomous drag-and-drop analysis. Because the data in the data business package is already linked together, this determines the degree of freedom of the data platform for immediate analysis. When analyzing the influencing factors of an indicator, you can choose any dimension to analyze the relationship between them. This determines the size of a factor's impact on the indicator.

650) this.width=650; "src=" http://img.blog.csdn.net/20160129144629668 "alt=" here write a picture describing "title=" "style=" border:none; "/ >

The Analysis Component component supports various styles of tables to present data with various business requirements. Includes lists, groups, and crosstab tables. The components support a variety of chart types including column, bar, Pie, area, combo, dashboard, and map. It is easy to create a series of interactive settings such as table conversion chart, add drillthrough, filter filter, add control, and so on page. You can switch to the chart with a single click by dragging and dropping the table generated by the indicator and dimension.

The platform supports a variety of charts and types are freely switchable. Supported icon types are column, column, line, stacked area, combo, bar, stacked bar, pie, dashboard, map.

650) this.width=650; "src=" http://img.blog.csdn.net/20160129144637155 "alt=" here write a picture describing "title=" "style=" border:none; "/ >

Summary Data Statistics various indicators support the summation of the average maximum, the minimum value and so on a series of statistical methods to provide choices.

the way metrics that support multiple calculation indicators can come from a field and can be calculated from a formula. In the calculation of the year, the chain, ranking only need to configure the interface to get the desired results. There is no need to pass complex formulas.

Data Early warning support data alert function for data in a certain data range can be a traffic light warning or data foreground warning.

multidimensional OLAP Analysis platform provides a variety of common OLAP analysis operations can be any multi-dimensional analysis drilling analysis, sorting, filtering and other analysis functions.

Any dimensional analysis platform provides data analysis of any dimension any dimension that needs to be analyzed can be arbitrarily added for the data to be analyzed. The chart setup process is similar to the need to note that most charts do not need to intentionally add analysis. Support for arbitrary dimension switching allows for free analysis of existing sample switching dimensions.

Multi-layer drilling because the relationship of dimension data is established when the cube is set up, the dimension can be drilled in multiple layers directly through grouping and hierarchical setting. Sets the data associated with the data between multiple drill-through settings to view the detailed values of the data through a multi-layer drill.

sort the sort based on the results of the query sort the dimensions according to the dimensions themselves sort the dimension according to the size of the summary metric to sort by the value of the formula. You can make ascending, descending, and custom sorts. Sorting data is selected to be sorted automatically according to the selected sort order method. Sorting is not affected by the sorting results after the pagination is displayed for global sorting.

Seven, technical characteristics

Key points of Data warehouse technology

• Dynamically generated bitmap indexing techniques handle types such as strings. NIO memory-mapped file technology quickly reads processing numeric types.

• Cube data storage support for offline use supports cube data timing full and incremental updates. Data processing mode of parallel computing for dynamic memory data cube technology.

• Fast packet filtering based on bitmap indexing enables multithreading operations to be non-intrusive. The bitmap index compression technique. Avoid the caching mechanism of repeated computations.

Data platform data modeling and data application flow

• Database generates a cube file the cube file establishes a data model based on the original data.

• When you access a design report, the bitmap index of the field that you want to use is preloaded to the memory-increase hit ratio.

• Process grouping using a bitmap index to process the data to be transformed generates the results required by using multithreaded grouping multithreading and memory-mapped files to generate summary results. And the result is set up a data cube model to avoid repeating calculation when the next fetch and partial fetch.

Data Platform Module

Analyze data associations

When the end user is analyzing the data, it is likely that the data is modeled when data is not linked together as a holistic view of the analysis and when dealing with such problems often requires the support of technical staff needs additional data modeling work platform to provide correlation settings based on the user's semantics and to correlate the data as long as you understand the semantics to get the required data 。

Analysis of influencing factors of index

An indicator or aggregated data is often affected by a number of factors such as sales will be affected by the product quality sales area time sales agent sales strategy competitive product prices and other factors, and when the final customer analysis needs to understand the overall impact factors. The previous BI tools were to add these analytic dimensions to the final presentation layer in advance to let leaders or business people choose such a problem with two unfavorable factors communication costs need to make it clear that the technical requirements of the business needs to modify the impact factors complex add Delete factors need to notify the technician. The platform's factor analysis directly targets the final analyst through optimized algorithms that provide all the impact factors and determine the key factors.

Viii. Summary of advantages

• Excellent processing performance for large data volumes

Data processing using automatic correlation between tables and manually establishing associations to achieve the relationship between data so that the data based on business relationships have a complete data structure. Users who understand the business can analyze data simply by selecting the appropriate data based on the business.

• Data platform data modeling and data application flow

• accesses the bitmap index of the field to be used to pre-load the design report to the memory increase hit ratio.

• processing the data by using a bitmap index to process a group the results required by the transformation generation and then using multithreading to group multi-threaded and memory-mapped files to generate summary results 。 And the result is set up a data cube model to avoid repeating calculation when the next fetch and partial fetch.  
  The grouping speed is fast. The non-interference between each grouping summary facilitates multi-threaded computing and distributed deployment optimization.

• Support Section compute the grouping rollup does not need to calculate all the values. The list speed is not limited by the amount of data.

• non-IT staff can perform instant analysis

The complexity of traditional bi is mainly embodied in two aspects

The first technician needs to spend a lot of time preparing the data. The underlying data for analysis is distributed in different places if you want this data to meet your business needs in a way that requires additional processing the data model that conforms to its tools is built according to the tools provided by traditional bi, and the process takes several months depending on the complexity of the business.

The second business personnel based on the data of some of the analysis requirements of the implementation process is complex. The traditional bi model is to pre-understand all the business requirements of leaders and business people and then prepare data for presentation based on these requirements analysis process when the decision-makers have additional ideas in the analysis process based on traditional design patterns they also need to communicate with technicians to prepare new data or design new analysis processes before they can get their own The analysis of the desired process also involves getting the technician to understand their needs, so the whole process seems to be quite complex.

The data Service module of the platform has the analysis design pattern and the indicator influence factor intelligent Analysis module can solve the above problems, so that the technician prepares the data without any code and complicated setup process to make it possible for non-IT staff to participate in development programming.

This article is from the "West Lake Bully" blog, please be sure to keep this source http://10143917.blog.51cto.com/10133917/1739859

Test run analysis Report of a company based on Finebi data decision platform