3

Today, when the trend of digital transformation is sweeping across major industries, more and more enterprises are beginning to pay attention to the deployment and application of BI (Business Intelligence) technology, expecting to obtain more business value from the growing data resources, thereby increasing profits and controlling risks. ,cut costs. BI can integrate, organize and analyze data, convert data into valuable information, provide support for enterprise management and decision-making, and become a decisive factor for enterprises to meet changes and business innovations.

Due to the importance of BI technology, enterprises hope to integrate BI capabilities in existing business platforms and systems on demand, so as to give full play to the advantages brought by data analysis in various scenarios and meet the increasingly diverse data analysis demands of enterprises. Deeply integrate BI capabilities with enterprise business. However, most of the common BI tools on the market are independent and packaged overall solutions, which are difficult to integrate with the front-end business system, and often fail to meet the needs in practice. In this context, embedded BI came into being.

The so-called embedded BI is to integrate various types of data analysis capabilities in the existing business system of the enterprise on demand. This kind of integration work generally needs to consider two points: on the one hand, it is essentially an upgrade process of the existing business system, and it is necessary to pay attention to the compatibility, stability, security and other indicators of the upgrade content and the original system; on the other hand, The business side generally hopes to deeply integrate professional data analysis components, rather than arbitrarily mounting a simple module to deal with it. These two points put forward higher requirements and challenges for the development team, which need to be taken seriously by the team.

Exploring the Architecture of Embedded Data Analysis Module

For many small and medium-sized enterprises, the software development team does not have the ability to independently develop In-House embedded BI solutions, and needs to seek the support of external third-party suppliers. There are also many specialized external vendors in the industry who have explored some market-proven best practices for embedded BI. We take the Wyn business intelligence embedded architecture developed by Grape City, a well-known enterprise in the industry as an example, to discuss how the data analysis module should be embedded in the existing business system:

As shown above, for business personnel, the data analysis dashboards, charts, designers, portals and other modules of the application function layer are the final effect of the embedded BI solution. Among them, the content and form of the module are generally determined according to business needs, such as integrating some sales data dynamic charts for a sales kanban to reflect the current sales situation in various regions in real time. Since business requirements are often diversified, the content and form of embedded modules are also very changeable, which requires the front-end technical layer to have stronger adaptability.

In the past, the front-end technology layer generally used the URL iFrame structure to implement module embedding. Now, more complex needs are more built with DIV solutions. In addition, taking Wyn business intelligence as an example, its BI module can also be integrated with common enterprise information systems such as Fanwei, UFIDA U8+, enterprise WeChat and DingTalk to enhance their data analysis capabilities.

The API layer is very important for embedded BI scenarios. For example, the API allows toolbars to be turned on and off based on user type, only displays specified data sources based on usage rules, and supports the creation of dashboards with different filters and options. Professional embedded BI can perform in-depth integration operations such as permission management, classification management, renaming, and deletion of dashboards/reports within the application software by calling APIs, and the interface between the application software and BI software is completely complete for end users. transparent. Of course, for simpler business requirements, it is acceptable to cancel the API layer in the embedded BI architecture, or only have a simple API layer.

Comparison of mainstream implementations

As mentioned above, for the development team, the key to the technical selection of the embedded BI solution lies in the choice of DIV and IFrame architecture, and whether to add a powerful API layer.

The IFrame architecture was very popular in the early embedded BI market, because of its simple principle, convenient implementation, and short development cycle, enabling enterprises to quickly realize the initial embedded BI requirements. However, although this method is simple, it has great limitations. For example, using IFrames makes it difficult to deeply integrate data analysis modules in the system. IFrame is more like the Flash element in the past, it is a relatively independent module. It is difficult to integrate and interact with other elements of the page, and even if it can be achieved, it requires a lot of effort and cost.

In contrast, the JavaScript-based DIV-level seamless embedding solution can utilize native JavaScript to integrate the entire dashboard into the project in the form of DIVs. Embedded chart elements have a highly open interface that enables data interaction with other elements. Even BI software as a whole can be embedded directly into existing systems through the DIV architecture, ensuring a seamless and intuitive user experience. Even if the current business requirements only stay at the simple diagram display level, considering the future upgrade and expansion potential, the development team is better to choose the DIV architecture to design the BI embedded solution.

On the other hand, the API layer can greatly simplify the operation of the embedded BI module by business personnel, which is often the functional goal that the development team needs to focus on. The embedded BI solution of Wyn business intelligence provides GraphQL API, and almost all interface operations can be simply completed through API calls. The following figure is an example of a simple API call:

![图片

GraphQL APIs do not require different URLs for different object operations. Different operations of different objects are called through a unified URL ( http://localhost:51980/api/graphql ), but the data submitted by various operations are different. It can be seen that the operation of the GraphQL API is very easy to use, which can greatly facilitate the development team and business team to meet various complex business needs. Let's take a look at the operation example of a data query API provided by Wyn Business Intelligence, from which you can experience the low threshold and convenience of the API:

图片

When we need to call a data set, we can complete the operation in two ways: POST or GET through this API. (示例URL 为http://10.32.5.7:51980/api/v1/datasource/b7d93485-66f2-4356-ada5-0a163579232d/query?queryType=sql&query=select *from Categories&token=27487CA0BE14CF599444E8553E5E07F78D5D1AB8646302A2715E4802FCB95F08&format=json;调用数据集的URL The format is POST/GET api/v1/dataset/{document id}/query.)

POST method, payload format:

 {
  "QueryType": "NONE|WAX",
  "Query": "some text like a WAX statement or empty"
  "DatasetParameters":{
    "Parameter1": "ParameterValue1"(单值),
    "Parameter2": "ParameterValue1,ParameterValue2" (多值使用逗号分隔)
  },
  "Format":Arrow | Json,
  "Options":{
  "Parameter1":"Value1"
  }
}

GET method, query parameters

 ?format=Arrow | Json
&$parameter1=value1
&$parameter2=value2_1
&$parameter2=value2_2
&option1=value1
&option2=value2

option1, option2 ... are option parameters, the prefix $ represents the dataset parameter, and multiple values are represented by repeating one parameter multiple times. The details of Option option parameters are as follows:

图片

图片

Only a few simple lines of code can complete the calling operation of the data set, which is undoubtedly very valuable for embedded BI scenarios. The API layer combined with DIV embedding can provide a satisfactory solution for embedded BI scenario requirements.

In general, although the iFrame architecture has certain advantages in terms of entry barriers, development costs and cycles, as the data analysis needs of enterprises become more complex, the DIV architecture can quickly show stronger scalability and adaptability. Teams can choose an iFrame implementation early on and migrate to a DIV solution as demand increases. At the same time, the development team often considers the implementation of the API layer at the beginning, which brings more convenience to the business team and lays a solid foundation for the later development work.

Embedded BI selection: how to choose the best solution for you?

When choosing an embedded BI solution, in addition to paying attention to the development cycle and development difficulty of the solution, the enterprise development team generally also considers factors such as pricing model, cloud support, and business system integration support:

  • cost. Not only the initial development cost, but also the long-term total cost of ownership should be considered, and future functional expansion, security enhancements, and other requirements should be included in the cost calculation.
  • Pricing Model. Embedded BI solutions provided by third parties often have multiple pricing models to choose from, such as pricing by user/server/CPU, or pricing by actual usage and usage time. Generally speaking, a relatively fixed pricing model is more beneficial to the enterprise user side.
  • Cloud support and business system integration support. Whether embedded BI can support public cloud, private cloud, local deployment, hybrid deployment and other modes is also an important consideration in the cloud computing era. In addition, the ability to integrate with other popular third-party enterprise systems (ERP, CRM, OA, etc.) can greatly expand the application scope of embedded BI solutions.
  • Authorization Management/Security. The embedded BI module often interacts with the confidential data of the enterprise, which requires the embedded BI solution to have good authorization management capabilities and security, so as to avoid the module itself becoming a weak link in the enterprise security line of defense, resulting in accidental leakage of sensitive data.

Considering the above factors, in practice, SMEs and development teams are more suitable for choosing mature third-party embedded BI solutions to meet their own needs. Taking the embedded BI solution provided by Wyn Business Intelligence as an example, it not only provides complete functions and convenient API integration based on GraphQL, but also supports powerful authorization management, enhanced security features, is compatible with multiple cloud deployment modes, and can be easily integrated to UF, enterprise WeChat and other systems. This solution can be embedded and deployed or used independently. It has good scalability and helps enterprises continue to deeply mine the value of data. To learn more about embedded BI solutions, you can click the link below for more information.

Reference link:
Wyn Embedded Business Intelligence: https://www.grapecity.com.cn/solutions/wyn
Data visualization large screen sample collection: https://www.grapecity.com.cn/solutions/wyn/demo
Embedded BI: https://www.grapecity.com.cn/solutions/wyn/embeddedBI
How to integrate with SaaS applications: https://www.grapecity.com.cn/solutions/wyn/saas


六一
556 声望347 粉丝

SegmentFault 新媒体运营


1