Discover what a Datamart is and its usefulness within a company. This subset of the Data Warehouse is intended for trades. This tool is at the beginning of Big Data.
Data warehouses are inextricably linked to Big Data. It is in these places that companies and suppliers store the files and their metadata in order to archive or process them. Usually known as the Data Warehouse, there are specific formats of data warehouses, subsets. Among these, we find Datamart.
Datamart vs data warehouse
A Datamart is a subset of the data warehouse that can be translated into French by data store or data counter. Here the goal is not to collect data before sorting them (the definition of the Data Warehouse) but to organize them according to business practices or targeted domains. They will serve to user groups in the company.
The Datamart brings together a set of organized, targeted, aggregated and grouped data to meet the needs of the trades. Technically, it is created from a relational database exploited from the SQL computer language and stored physically on a hard disk through a database management system.
Datamart definition: two schools of thought
Its principles and operation are theorized by two computer science specialists: Bill Inmon and Ralph Kimball.
Bill Inmon, considered by many as the creator of the Data Warehouse, this researcher has written more than 40 books and more than 1000 articles on this subject. It defines Datamart as a data stream from the Data Warehouse. It groups the specialized, aggregated data for a particular job functionally. In this approach, it is not at the heart of the data warehouse, but at the periphery of the data warehouse.
Ralph Kimball, also a computer scientist and entrepreneur who has written many books on data warehousing. According to him, the Datamart is a subset of the Data Warehouse composed of detailed or aggregated tables, linked together. The idea is to make it accessible, fast and representative of an activity in a company. According to Kimball, Datamart is the Data Warehouse.
These two conceptions of Datamart converge towards the same appreciation of this item. It is an extract of all the data of a company; it contains only the necessary. Unnecessary data is removed and the logging is done according to the user’s request. Thus, the trades are not disturbed by parasitic or contiguous data. Two Datamart usage patterns prevail Inventory Inquiry (or Business Activity History), or Stream Activity (Live Orders).
The architecture built on a relational database and the need to silo the data facilitates the speed of access but locks the possibilities. An unusual request made from a database query software, for example, a reporting tool, will not give the expected result unless a modification of access to other parts of the Data Warehouse.
Several uses of this type of database are possible. Three professions are particularly concerned: marketing, commerce and human resources.
Datamart marketing and commercial
This type of relational database focuses on the needs of marketers whose job is to identify prospects and target customers. With such a tool within their reach, they can consult all the contacts registered with the company. Names, first names, phone numbers, physical addresses and e-mails are some of the information that can be consulted. With an associated reporting tool, he can quickly know who he has contacted, whether he has managed to sell a product to him if he is satisfied, and so on. Everything can be associated with the impact of behavior on the company’s turnover.
Datamart Human Resource
Here we use the employee information to record arrivals and departures of employees, the average age of employees, professions represented, seniority in the company, compensation, etc. This provides statistics and reports that facilitate hiring or firing decisions.
This time, it’s about providing business intelligence about the financial and administrative health of the business. The overall turnover, by sector, as well as the costs, are analyzed. Likewise, audit elements such as invoices, their origins and purchase orders. It can identify the unpaid, the profits, the payroll of the company according to the data flows used.
Datamart VS Cube
New Business Intelligence users often confuse Datamart and Cube. They are easy to understand; they are two elements relating to business applications within a Data Warehouse. Only the first is a subset of this infrastructure and gathers all the data about an activity. The Cube, for its part, makes it possible to carry out queries to answer the particular questions of the trades. What is the age of a worker population? What is the turnover rate? That’s what this tool will do.