A data warehouse is a central data store in a company or organization that collects relational data from various sources. The information is transferred from different transactional systems or other relational databases to the data warehouse, where it is available to analysts and decision-makers.
What is the function of a Data Warehouse?
The data warehouse is used in many areas of the business environment. The database is used throughout the company to make data-driven decisions or to examine processes. Since the central data warehouse draws information from many different systems, it is seen as a single point of truth. This is to ensure that everyone in the company is talking about the same data and that decisions are based on this information.
Across departments, the data warehouse can be used for the following tasks:
- Cost and resource analysis
- Analysis of internal processes (e.g. production, hiring, etc.)
- Business Intelligence
- Calculation and provision of company-wide key performance indicators
- The data source for analyses or data mining
- Standardization of company-wide data into a fixed schema
What properties does it have?
When creating central data warehouses, one can be guided by certain characteristics that should help to better narrow down the structure and the necessary data of the warehouse.
A data warehouse contains information on a specific topic and not on individual business transactions. These topics can be, for example, sales, purchasing, or marketing.
The warehouse aims to support decision-making with the help of business intelligence and targeted KPIs. This interpretation is also supported by the fact that information that is not relevant to decision-making or used for analysis does not end up in this central database in the first place.
The warehouse integrates data from a wide variety of systems and sources. Therefore, a common schema must be created for the information to be uniform and comparable. Otherwise, central analysis and KPI creation are not possible.
Time Period Reference
The data warehouse stores data from a specific time period and is therefore related to the past. Furthermore, the data is usually transmitted in aggregated form, for example at the daily level, so that the volume of data remains limited. Thus, the granularity may not be fine enough, as one is used to the operational systems.
On the other hand, the operational systems are time-based, as they output the current information accumulating. At the same time, the information can be viewed in great detail.
Another important feature of central warehouses is the non-volatility of the data. In operational systems, information is usually only stored temporarily for a short period of time, and as soon as new data is added, old data is overwritten. In a data warehouse, on the other hand, data is stored permanently and old data persists even when newer data is added.
What are the Advantages of a Data Warehouse?
By establishing a central data repository, more people can access information more easily and use it in the decision-making process. As a result, more rational decisions are made that do not depend solely on the opinions of individuals.
Furthermore, it offers the possibility of combining information from different sources and thus creating new ways of looking at things. Previously, this was not possible due to the physical separation of the systems. It may be interesting for a company to combine data from the web tracking tool, such as Google Analytics, and sales data from the sales tool, such as Microsoft Navision, and evaluate them together. This allows the performance of the e-commerce store to be evaluated and the user journey to be viewed together with the actual purchase.
Data quality is easier to ensure with centralized storage since only one system and the data it contains need to be assessed. Fragmented data sources distributed throughout the organization, on the other hand, are very difficult to keep track of. The same is true for controlling access by individuals. When data is distributed throughout the organization and may not even have been stored in databases, it’s hard to track who has access to data they shouldn’t normally see.
Which other components belong to the Data Warehouse?
In addition to the data warehouse itself, other systems and components must be added to an architecture to enable centralized data storage. These include:
- ETL tool: To ensure querying and consolidation from various data sources, a tool is needed that queries the sources at regular intervals, consolidates the data, i.e., brings it into a uniform schema, and finally stores it in the data warehouse.
- Metadata storage: In addition to the actual data, so-called metadata must be stored. This includes, for example, user rights, which define which changes a person may make or which tables may be viewed. In addition, the data and table structure are also recorded, i.e. how the tables are related to each other.
- Further tools: The existence of the data in the data warehouse alone does not really help. In addition, other tools must be built in to use and process the information from it. Often, for example, a suitable visualization tool, such as Microsoft Power BI, is added so that data analyses can be created and aggregations performed. These ensure that knowledge can really be generated from the data.
What are the differences to the Data Lake?
This central database can additionally be supplemented by a data lake, in which unstructured raw data is stored temporarily at a low cost so that it can be used at a later date. The two concepts differ primarily in the data they store and the way the information is stored.
|Features||Data Warehouse||Data Lake|
|Data||Relational data from productive systems or other databases.||All Data Types (structured, semi-structured, unstructured).|
|Data Schema||Can be scheduled either before the data warehouse is created or only during the analysis (schema-on-write or schema-on-read)||Exclusively at the time of analysis (schema-on-read)|
|Query||With local memory very fast query results||– Decoupling of calculations and memory|
– Fast query results with inexpensive memory
|Data Quality||– Pre-processed data from different sources|
– Single point of truth
|– Raw data|
– Processed and unprocessed
|Applications||Business intelligence and graphical preparation of data||Artificial Intelligence, Analytics, Business Intelligence, Big Data|
This is what you should take with you
- A data warehouse stores company-wide information centrally.
- This is intended to support data-driven decisions and make business intelligence possible.
- The unstructured raw data in the data lake provides an excellent complement to the relational and processed data in the data warehouse.
Other Articles on the Topic of Data Warehouses
- Amazon Web Services provides a good summary here.