Alan Simon

将数据建模视为制造公司的数据和系统,以建立汽车,家用电器或飞机的模型,然后才能构建实际物体。模型帮助我们描绘了不同对象和子系统如何相互关系的概念图片。我们可以在昂贵的时间和资源投资到实际对象的昂贵投资之前就可以解决这些细节。

在数据模型的情况下,我们可以列出主要数据将如何相互关系,它们各自的特征和属性是什么以及哪些业务规则适用于整个模型以及较小的组件。数据科学家,数据分析师,软件开发人员和其他人可以使用数据模型的描述性内容清楚地了解每个数据在构建分析模型,软件数据定义以及组织数据的其他用途之前的实际含义。

Mastering Data Modeling Fundamentals

Last Updated March 2020

Bestseller
  • 37个讲座
  • Beginner Level
4.4 (3,305)

Entity-Relationship Techniques and Best Practices |艾伦·西蒙(Alan Simon)

Explore Course

In this article, we’ll dig deeper into the fundamentals of data modeling and answer common questions of why it’s an important piece of data science.

数据建模如何工作?

数据建模有助于我们以高度概念的方式开始处理数据。它的目的是比许多数据库更近地反映现实世界。此外,建模将沿概念性物理生命周期转换和映射数据,一直到高度调谐的数据结构,旨在支持最佳的在线性能和响应时间。

In other words, data modeling is a discipline that mirrors other types of technology that begin with conceptual views and end with technical implementations — like software development and business process modeling, for example.

数据科学家什么时候使用建模?

数据科学团队负责构建复杂的分析模型,以基于广泛的预测和探索/发现分析。这些分析模型的关键构建块是正确的数据。这意味着准确地准确地了解每个数据元素的含义,其内容的可能范围或值列表,是否必须存在数据实例或是否允许丢失数据(例如,如果允许零值),数据与其他数据之间的关系数据等等。

Data models help the data science teams fully understand the data and identify any possible issues that could impact the validity and accuracy of their models.

Mastering Data Modeling Fundamentals
Alan Simon
4.4 (3,305)
Bestseller
初学者的数据仓库基础知识
Alan Simon
4.5 (7,858)
Bestseller

Courses by Alan Simon

三种类型的数据建模是什么?

现在,我们知道数据建模的工作原理的基础知识,让我们进入三种类型的数据模型,它们在数据库系统中都有自己的角色。

Conceptual data model

概念数据模型理想地可以尽可能地反映现实世界。换句话说,我们试图避免特定数据库管理系统的约束,这通常需要以非常特定的方式构造我们的数据,有时包括“人造”数据库表或克服这些约束所需的其他结构。取而代之的是,我们希望概念模型代表特定的业务问题,或者我们正在为其构建的任何东西,以及实际的对象,特征,关系和业务规则。

逻辑数据模型

A logical data model maps our conceptual model into a database-specific set of structures. For example, in a relational database, we have certain constructs like tables comprised of rows and columns. We take the model that we built at the conceptual level and apply whatever transformations we need to make in order to make the model now conform to the rules of relational databases, or whatever type of data management system we’re using.

Physical data model

At the physical level, we make adjustments to the logical model in two ways. The first way is driven by product-specific features and limitations. The second way attempts to increase performance and response time once the database is actually filled with data and implemented.

这三个级别都至关重要,并且数据建模生命周期理想的是在这三个级别中的每个级别中都有实质性参与。

什么是常见的数据建模工具?

A data modeling tool is a specialized piece of software that not only permits us to draw major data modeling constructs — entities, attributes, and relationships — but also attaches semantics to those objects. These meanings might be as simple as the data types and sizes of attributes, or as complex as the business rules that govern particular relationships between two entities on a data model.

The semantics then help us translate and convert highly conceptual models that mirror the real world into database-specific constructs that become the data management platform for one or more applications and systems.

Data modeling tools can range from general-purpose packages such as Microsoft Visio, which has evolved into a full-functioned product to build and manage data models, to specialized software such as CA Erwin, ER/Studio Data Architect, and many others. Here’s a look at five of the most popular tools.

  1. Microsoft Visio

如果你不需要非常专业的软件,then Microsoft Visio might be a great place to start. Its diagrams are professional and intuitive. Additionally, the flowchart templates are modern and you can collaborate with your team in the tool to get real-world insights.

  1. Ca Erwin

如果你使用Windows操作系统和基于“增大化现实”技术e looking for a data modeler to work with structured and unstructured data, Erwin Data Modeler is an award-winning, high-quality modeling tool.

  1. ER/Studio

Idera的ER/Studio是另一个最适合Windows的软件。使用企业团队版,您可以对BI,大数据,ETL,NOSQL和关系来源进行建模数据,并实现命名标准,定义词汇表术语,并在存储库中保存模型,以供其他团队成员查看。

  1. PowerDesigner

SAP PowerDesigner工具最好用于数据建模,LYNC和同步和元数据管理。该界面易于导航,并具有存储库,因此您可以轻松地与团队合作。该工具最适合Windows。

  1. Navicat

Use Navicat across any of the operating systems. It’s a more cost-effective way to build high-quality data models than some of the other tools, and you can still perform many of the same tasks. Navicat is best for forward and/or reverse engineering, importing models from other data sources, or generating complex SQL/DDL.

流行的数据建模技术

There’s more than one approach to data modeling. The right technique will depend on your use case and business requirements. Some data modeling methodologies can be thought of as database-centric, which means they are dedicated to graphical depictions of data and the relationships among data objects, with the eventual goal of designing and creating a database. Other techniques are more holistic in nature, addressing data models in concert and alongside process modeling and other aspects of systems development.

Most approaches trace their roots back to entity-relationship (E-R) modeling, which began in 1976 with Peter Chen’s academic paper and has evolved over the years. Another technique closely related to classic E-R modeling is the so-called “crow’s foot notation” which follows many of the same rules as “classic” E-R modeling but uses different symbols to represent data and business rules.

The E-R model, regardless of notation, represents real-world objects. An entity is a thing in the real world. It has a set of properties or attributes and those properties can have values. Because there can be many entities, those entities can have relationships with each other. E-R modeling, therefore, models those entities and the attributes and relationships among them.

现在,我们已经介绍了数据建模,因此您可能想知道它如何适合您的职业发展。必威真人投注IT和数据专业人员可以将数据建模专业作为以数据为中心工作的一部分(他们所做的)(数据仓库或大数据,例如)或与数据模型一起工作,这是更大的作品的一部分,例如自定义软件开发,包装软件开发或数据科学和机器学习。必威真人投注通常,对于任何以任何身份使用数据的人来说,数据建模都是至关重要的技能。有人可以使数据建模其职业生涯的中心,也可以成为更大技能工具包的一部分。

赋予您的团队能力。领导行业。

Get a subscription to a library of online courses and digital learning tools for your organization with Udemy for Business.

请求演示