Two of the core elements of data are the Entity and Attribute.
An entity is a uniquely identifiable group (or class) of persons, things or concepts for which the business area wants or needs to keep information. Entities can represent abstract concepts like events or classifications.
Examples: Employee, Job Classification, Paycheck
An attribute is a characteristic that further describes an entity. It defines a particular piece of information associated with an entity. Each attribute belongs in one and only one entity.
Examples: Employee number, Employee first name, Employee Last Name
Attributes should be grouped by the entity that they describe and should be defined somewhere. We recommend using this Attribute Metadata Template, but other options include defining them in a glossary or data dictionary.
Understanding the metadata associated with your attributes (data elements) is key for:
- Establishing a shared understanding of the meaning of your business data (definitions of terms)
- Helping elicit the rules around the data from the business (what does the data look like? Is it mandatory or optional? Is it unique? Are there valid values or ranges we need to adhere to?)
- Ensuring that data is broken down appropriately. For example, what data elements make up an address? Or a phone number? Or a name?
- Providing a great basis for a data architect to build out a physical data architecture.
Use this Attributed Metadata Template to Identify:
Name: single attribute or combination of attributes that is chosen to uniquely identify every instance of the entity.
Unique (U): Identify if the attribute is unique or non-unique. Examples:
- Unique: Employee number
- No-Unique: Last Name
Mandatory (M): A mandatory attribute indicates that every instance of the entity must have a value. Examples:
- Is a first name required for every employee? YES
- Is a last name required for every employee? YES
- Is a middle name required for every employee?NO
Repeating (R): If an attribute can have more than one value, it is called a repeating attribute. Finding a repeating attribute is a good indicator that further analysis needs to be done.
- Example: Employee street address “repeats” because an employee may have a work address, home address, vacation address, etc.
- The repeating attribute should be fully resolved so that we understand all the variations of the data.
Data Type: A data type is typically defined as “a standard form of data”, or a standard for how the data looks and behaves. Types include:
- String: This type is defined as being any printable character, i.e. “A”-”Z”, “a”-”z”, “0”-“9”, plus all symbols and the space character. This type is also known as character, character string and alpha-numeric.
- Number: This type can be used in calculations. There are many number types, examples include an integer or decimal.
- Date/Time: This is a special number type that is interpreted by the system as a date and/or time.
- Other: Any data that doesn’t fit the other 3 types listed above – may be used for video, audio files, graphics, etc.
Valid Values: Some attribute values are limited to a specific set of values or numeric range. If so, document this information in your template.
Tip! When an attribute has a long list of valid values, or a list that may be used by multiple attributes, document them in a separate document or template.
Default Value: If the attribute should default to a particular value, indicate that value here.
Owner: Person or department that is responsible for the data integrity.
Definition: Attribute definitions do not need to be long; sometimes a single sentence will suffice if the entity is well defined and a meaningful name has been chosen.
An attribute definition should:
- Include an example of a single instance of the attribute
- Describe how a value is assigned to the attribute, example: employee numbers are assigned sequentially.
There may be other metadata items you need to add depending on the intent, for example: