Summary: in this article, we will discuss the data mart concept, different types of data marts including stand-alone data marts and dependent data marts.
What is a Data Mart?
A data mart is a set of subject areas organized for decision-making support based on the specific needs of a group of business users or departments.
There are two types of data marts: independent or stand-alone data mart and dependent data mart.
Stand-alone data mart
A Stand-alone data mart focuses exclusively on one subject area and it is not designed in an enterprise context. For example, manufacturing has its data mart, and also human resources do and so on. stand-alone data mart gets data from multiple transaction systems in one subject area or department to support specific business needs. a stand-alone data mart may use a dimensional design or entity-relationship model. Analytic or business intelligence tools query data directly from data mart and present information to users. The picture below is a typical Stand-alone data mart.
A stand-alone data mart takes a very short time to build and bring the visible result to specific departments with less cost. However, if you look at the whole system landscape where multiple data marts exist, you will see that different ETL tools need to build for different transaction systems in different technologies and the data is duplicated in several data marts. From the business perspective, each data mart is built to address a set of specific business needs, what if the needs expand? And what if you want to analyze data across functions or departments? The inconsistent data, such as the definition of a product, will make the information comparison between departments impossible.
Dependent data mart
According to Bill Inmon, a dependent data mart is a place where its data comes from a data warehouse. Data in a data warehouse is aggregated, restructured, and summarized when it passes into the dependent data mart. The architecture of a dependent data mart is as follows:
There are several benefits of building a dependent data mart:
- Performance: when the performance of a data warehouse becomes an issue, build one or two dependent data marts can solve the problem. Because the data processing is performed outside the data warehouse.
- Security: by putting data outside data warehouse independent data marts, each department owns their data and has complete control over their data.
- KPI tracking: dependent data marts are ideal places for building and tracking KPIs over a long period of time.
In this article, we’ve examined data mart definition and discussed stand-alone data mart and independent data mart architecture.