Semantics in general means the meaning of a message behind the words, so when we say semantic in terms of BI, it means the meaning of data from the user's perspective. It is the keystone of any data warehouse project, as it allows better understanding of requirements, the design of successive data models, and a link between the physical data model and the reporting tool, in our case MicroStrategy. In the previous chapter, we learned about setting up an environment for MicroStrategy, so in this chapter we can dig deeper and understand about the semantic layer and the objects needed to present data to an end user.
This chapter will cover:
- BI architecture
- The project design process
- Defining schema objects
- Creating attributes
- Defining parent-child relationships
- Building hierarchies
- Creating facts
- Defining public objects
- Creating simple and complex metrics
- Adding filters
- Creating prompts
BI is all about delivering timely and relevant information to the right level of audience, which helps them to make an appropriate decision. In this section, we will learn about BI architecture and the components involved in it.
Source system: This is any stored data that will be used for BI projects. It could be a database, mainframe system or flat files, or any other system that stores Online Transaction Processing (OLTP) data. This is where data is stored in its raw form.
ETL: This process then pulls the data from the source system and prepares it for the data warehouse. It involves three steps:
- Extract: The process of reading data from source systems
- Transform: The process of converting data from its previous form to the form it needs to be in, either by combining data from multiple sources or by applying business rules and logic
- Load: The process of writing data into the target database, that is, the data warehouse
Data warehouse: This is typically an Online Analytical Processing (OLAP) system. It is the way to store data to be used in the analysis. Data can be stored in regular databases such as MS SQL Server, Oracle, and so on.
BI Platform:In our case, this is MicroStrategy. Some of the main components of MicroStrategy are:
- Metadata: This is data about data. What that means is that it contains information that facilitates the retrieval of data from the data warehouse using the MicroStrategy application. It stores MicroStrategy object definitions and maps them to the data warehouse's content. The MicroStrategy application uses metadata to translate user requests into SQL queries, and the results from SQL queries into MicroStrategy objects. Therefore, we can say metadata acts as a central repository for all the object definitions, such as facts, attributes, filters, and so on.
- MicroStrategy applications: These provide the ability to present the data in a superior form for analysis. Users can create reports, grids, graphs, and dashboards, and access them via MicroStrategy Web, Desktop, and Mobile, and MS Office too. It supports in-memory analytics with intelligent cubes. In addition to this, users can use a software development kit (SDK) to customize the application and integrate with other applications.
- MicroStrategy Architect: This project design tool allows users to define all the required components to build a project from a centralized interface.
At this point, we know about all the components, but how these components come together is the next thing we will learn in the project design process.