"8 Questions" about BI Business Intelligence | One article to understand big data BI

The concept of business intelligence is no longer explained here. Regarding BI, from the past understanding, search and some questions and answers in Zhihu, everyone's confusion mainly focuses on the relationship between big data and BI, some technical issues of BI, and the BI industry. and personal career prospects. Here are 8 questions, each of which has been carefully answered, hoping to bring help to everyone.

Question 1 : What is the relationship between business intelligence BI and big data, and how to choose?

BI (Business Intelligence) is business intelligence , which is a complete set of solutions to effectively integrate existing data in the enterprise, provide reports quickly and accurately, and propose decision-making basis to help enterprises make wise business decisions. .

Big data refers to the collection of data that is captured, managed and processed with conventional software tools within an affordable time frame. It requires new processing modes to have stronger decision-making, insight and process optimization capabilities Adapt to massive, high growth rate and diverse information assets.

The two are different concepts. Compared with big data, BI is more inclined to decision-making and is suitable for supporting business indicators. Big data has a broader connotation, tends to describe individuals, and is more about personalized decision-making.

Today, large Internet companies use a big data architecture such as hadoop—data warehouse—reports developed by themselves, OLAP analysis, etc., or use mature business intelligence reports and BI analysis software at the front end. Big data BI such as FineBI Can connect with big data platform. Traditional enterprises and small companies do not have so many business analysis needs. Most of them seek simple reporting functions of excel, business systems or professional reporting tools to solve problems. Generally, large and medium-sized enterprises will build data warehouses when the amount of data is large, and use BI to solve problems. Front-end analysis display. Of course, many traditional enterprises use big data technology for specific businesses (such as user portraits, risk control analysis).

Generally speaking, the choice of big data or BI depends on the needs. Most of the big data components are open source and require a lot of human development. BI is mostly commercial, and requires a certain amount of capital and a certain period of time to implement the project.

Question 2 : What is Big Data BI ? What is Self-Service BI ? How is it different from traditional BI ?

Big data BI

Big data BI has the following characteristics: it can connect to big data platforms such as hadoop for data analysis; it can process large data volumes (over 100 million) and has fast response speed. The test is the data processing computing performance and database of BI. performance.

Tradition BI & Self-help BI

Traditional BI focuses on building a data platform, providing report services, and is dominated by IT; self-service BI focuses on data analysis and is dominated by business analysis. The data processing flow of the two is the same.

Traditional BI: usually refers to a large and comprehensive unified report or analysis platform within an enterprise. Representative veteran BI tool manufacturers such as IBM's cognos, Oracle's OBIEE, SAP's BO, etc. all contain rich functional modules, which are more suitable for building integration A large and comprehensive unified platform. Traditional BI is generally aimed at IT R&D personnel, who are mostly concentrated in the technical department of the enterprise. The traditional BI building method is basically as follows:

关于BI商业智能的“8大问”|一文读懂大数据BI

Self-service BI : It is aimed at business analysts without an IT background. Compared with traditional BI, it is more flexible and easy to use, and to a certain extent, it gets rid of the heavy dependence on the IT department. Different from the previous "IT-led reporting model", it has turned to a "business-led self-service analysis model". Common application scenarios of self-service BI:

关于BI商业智能的“8大问”|一文读懂大数据BI

Question 3 : What is the multidimensional data model and OLAP of BI, and where is the practical value?

Just imagine the scenarios when analyzing business data, and often look at business metrics from different perspectives. For example, when analyzing sales data, factors such as time period, product category, distribution channel, geographic distribution, and customer group may be considered. Although these analysis angles can be reflected by reports, each analysis angle generates a report, and different combinations of each analysis angle generate different reports. Every time an analysis is attempted, a number must be drawn, which will make the IT staff’s The workload is quite large.

The role of OLAP is to prepare all dimensional conditions and aggregated values ​​as much as possible, so that users can analyze according to any dimension during analysis.

以BI的实际应用来讲,拿到数据,可能需要下钻到比较粗的粒度观察数据,比如从日期时间维度、从地域品类维度来分析数据,对应到BI的操作上,就是拖拽维度、过滤排序、维度切换,钻取等操作,cube或者数据仓库就要响应这种操作,这就使用到了下钻、切边、切块、转轴等功能。

问题4:商业智能BI在数据分析工作中的作用,是必要的吗?

在数据分析过程中,BI也算是一个工具,能自助取数,用于快速分析,制作分析报表。很多互联网、零售、金融企业会有自己的数据分析团队会专业的分析人员,使用的工具可能从SPSS、SAS、R、Python不等,这些工具能对准备好的数据做数理统计分析,取数的工作大多还是要交给IT人员去做。像目前的自助式BI因为上手很简单,对于多维度的数据可以从各方面来展示,而且能及时生成数据报告,可在平台上管理报表和分析表单。所以是否有必要,因需求而异。

关于BI商业智能的“8大问”|一文读懂大数据BI

问题5:BI如何选型,需要考虑哪些点?

BI工具可分为传统型BI以及自助型BI。传统型BI,国外以SAP BO、cognos、Oracle BIEE等为主;自助型BI,比如国外的Tableau、Qlikview,国内的FineBI、永洪bi等等。

站在产品的企业的角度,可以从领先能力、产品能力、服务能力以及价格能力去着手衡量。可通过海比研究给出的一套《BI选型指标体系》来判断。

关于BI商业智能的“8大问”|一文读懂大数据BI

1、领先能力=行业地位+领先性

比如公司在行业中的低位、市场占有率、公司在该领域的专注性以及技术的领先性。商业智能目前的市场格局不算大,可扩展到报表领域去衡量。

2、产品能力=公司产品线+核心产品功能+解决方案

一般来讲,公司的产品线越完整,相关产品的整合能力越强,越好。但是,最重要的还是产品的功能是否实能解决企业最关注的的问题,是否能覆盖更多行业,BI解决的是行业通用性的问题,解决能力越强,产品越优秀。

3、服务能力=服务专业能力+维护能力

BI的实施很考察人员的专业性,过去由于国外IT巨头的称霸,很多产品的项目都承包给第三方实施,造成服务脱节。现在很多涌现的国内软件公司一般都会有专业的实施团队,本地化服务很占优势,所以这一点不妨考虑本土产品。

4、价值能力=成功案例+性价比

选型前可看看同行业的企业伙伴们用的是什么类的BI工具,使用情况如何。包括从功能费用、项目实施费用综合考虑来看的性价比。

问题6:如何实施BI

实施BI的前提,最重要的是基础数据的统一。比如货品信息,客户信息,公司内部信息。缺少的数据虽然可以临时补,但是随着公司业务的扩展,这种数据化运营的方式需要不断精细,数据管理的规范任务要落实到业务员的考察,如果得到领导的支持会更容易推动。有了这些齐全的数据,BI的实施才有保障。

然后是业务的统一。比如销售模式,采购模式,结算方法,质量管理的统一。比如销售模式不统一,有的分公司先结算后配送,有的公司先配送后结算,业务形式不统一,口径不统一,就会造成数据的时间差。

其次是业务部署。每个公司的业务部署不同,有的是集中部署有的是分销部署,如果BI是放在总部实施,需要将各地分散的数据统一起来,建立数据仓库,保持基础数据的统一,但其中,如何提高速度,如何优化配合方式,这点需要研究。

BI人才储备是否足够,需要业务人员和信息人员的积极配合,这个效果才能够比较良好的推动,而且还能够持续的发展。为了让技术和业务人员更好地贴合,要将技术和业务有效结合,最大效率的把报表和BI系统的功能发挥出来。

对于上BI,还有其他考虑,比如价格预算,比如是否用开源,比如后续开发和维护,这里做个统一的解释。

明确业务需求:强烈的业务需求,明晰的业务目标,能否抓住核心是一个项目成败的关键。

“产品+定制+服务”的建设思路:是否要选择开源的产品?如果你有很强的开发能力,可以考虑。但建议专业的事情还是交给专业的工具来做,传统企业不比互联网企业,互联网企业是以数据来驱动的,与传统企业的模式不一样,再说,后续维护也是成本啊。业务项目建设如果不借助比较成熟的产品工具,从技术代码进行创新式的开发,不经过迭代以及检验很难规避风险,很难形成一个成熟的产品。如果觉得国外BI产品“庞大”,完全可以选择FineBI这一类轻量化的工具。

最后,就是认清技术力量的现状,不妨建议敏捷开发、迭代开发和重构,注重技术和管理的配合。

问题7:做BI人的前景在哪里?

刚入门BI这一领域的人,未来的职业发展可以走技术、走管理、走开发。

1、走技术方向:(按照技术路线进行划分)

ETL,这块是BI永恒的重点之一,需求也是一直持续,只是相对来说,ETL会比较枯燥。在这一块,掌握一两款顺应潮流的大工具,拥有相应年限的工作经历,行业性要求不太高,可以找到一个不错的岗位。DS、INFA、SSIS这些都是蛮有需求的。

数据仓库,主要指的数据仓库设计,架构设计等,一般来说LEVEL会比较高,薪水待遇也还行,属于偏高端人才了,一般都会要有5年、7年或更多年限的经验,对行业性经验要求比较高。

前端应用,SAP BO、COGNOS、BIEE等工具的熟练应用,可以做甲方内部顾问也可以做乙方项目顾问。从前端切入去接触到更多业务和需求,对提高自己的业务水平有好处。

数据挖掘数据分析方面,这块个人认为是最有前景的,数据分析师的需求一直在增加,但和BI的背景不是非常贴合,要学数据统计学知识,R、Python语言等,学的东西很多。

2、走管理路线

管团队管某块业务,做项目总监,而后上升到CIO之类的,偏向管理属性,对人的沟通交流尤其是与高层交流的水平较高,比较IT在企业大多属于业务支撑部门,很多事情很难推动,同时还要思考如何提升IT在企业的地位,这个你只要观察自己的部门领导怎么做的,慢慢摸索了。

3、走开发

第三方软件开发公司了,比如SAP之类的公司(有点难度),或者国内帆软、永洪等其他BI公司的软件开发了,难度是要有一定的程序开发基础,但是对业务的理解也能带来一些帮助。再或者是去一些创业公司带团队做BI产品,现在做前端可视化分析的公司有很多,虽然不完全类同于BI,但有很多共通之处。

问题8:如何系统的学习BI知识?

这里给出一个学习框架

1、学习数据库知识, 掌握基础技能SQL

2、技术选择:数据仓库 / ETL / 前端开发等等

3、 选择技术工具:

数据仓库-Oracle、SAP HANA、Hadoop都是主流;

ETL- informatica 、kettle;

自助式BI工具-Taleau、帆软FineBI、Power BI

4、学习业务知识

5、实操数据分析工作

Guess you like

Origin http://10.200.1.11:23101/article/api/json?id=326688709&siteId=291194637
Recommended