Data Scheme is a diagrammatic representation of the structure of data. It represents any set of data that is being captured, manipulated, stored, retrieved, transmitted, or displayed.  A Data Scheme can be a complex diagram with all sorts of geometric figures illustrating data structure and data relationships to one another in the relational database within the data warehouse.

As an example, let us take a generic website and illustrate the Data Scheme.

One of the general data categories in website Data Scheme is the User Accounts and Privileges and Watchlist. The Data Scheme may draw one big box for User Accounts and Privileges and Watchlist. Within this big categories are four smaller data categories boxes named User, Watchlist, User Group and User New Talk.

The User box contains basic account and information about users such name, password, preferences, settings, email address and others. The Watchlist box contains registered users and the pages the user watches, the namespace number, notification timestamp and others. The User Group permission box maps users to their groups with defined privileges. The User New Talk box stores notification of user talk page changes for the display of "You have have new messages" box.

Within each of the four boxes are defined the data and each of the data names as well as data types. The Watchlist box may contain the following data, the corresponding name and type:

user_id: INTEGER (5)
user_fullname: VARCHAR (255)
user_password: TINYBLOB
user_email: TINYTEXT
user_options: BLOB
user_token: CHAR (32)

The same structure goes with the other tables as well. It is very clear that all data structures are being defined with names and data types. In a real data scheme diagram, there could be hundreds of boxes, data names and types and crossing lines connecting one entity to another.

The graphical look for a data scheme has some similarities to a flowchart which is a schematic representation of an algorithm or a process. But while a flowchart allows business analysts and programmers to locate the responsibility for performing an action or making correct decisions and allowing the relationship between different organizational units with responsibility over a single process, a data scheme is just merely a graphical representation of data structure. There is no mention of any process whatsoever.

It may be a good point to note the Data Scheme may or may not represent the real lay out of the database but just a structural representation of the physical database. To a certain degree, the data scheme is a graphical representation of the logical schema in data modeling. A logical schema is technically a data model that describes semantics as data are represented by a particular data manipulation technology.

Data Schemes are handy guides for database and data warehouse implementation. They can be compared to an architect’s blue print of a house or a building wherein it is easy to locate some key points without spending too much time digging deep into minute details. Because of its graphical nature, professionals implementing a data warehouse will not have to strain their eyes on data structures and focus more on other details especially in dealing with programmatic codes.

Data Schemes are also highly useful in troubleshooting databases. If some points of the database are faulty, Data Schemes helps to pinpoint the cause of the error. Some errors in database and computer programming languages which are not related to syntax can be very hard to trace. Logic errors and errors related to data can be very hard to pin down, but with the help of a graphical Data Scheme, errors may be made easier to spot.