Metrics

Metrics specify the way in which data is measured. This is used for reporting purposes to show which types of exceptions occur in your data. For example, if the condition is designed to evaluate the record's completeness (meaning, for example, that all addresses contain postal codes) then you could specify "Completeness" as the data quality metric.

You can select one of the predefined metrics listed below or specify your own metric by clicking the Add item button and completing the fields as necessary. You can also edit metrics by selecting a metric, clicking the Edit item button, and making any necessary changes. You can also filter the list of metrics shown by entering search data in the Filter field. The results will update dynamically.

• Accuracy—The condition measures whether the data could be verified against a trusted source. For example, if an address could not be verified using data from the postal authority, it could be considered to be an exception because it is not accurate.
• Completeness—The condition measures whether data is missing essential attributes. For example, an address that is missing the postal code, or an account that is missing a contact name.
• Consistency—The condition measures whether the data is consistent between multiple systems. For example if your customer data system uses gender codes of M and F, but the data you are processing has gender codes of 0 and 1, the data could be considered to have consistency problems.
• Interpretability—The condition measures whether data is correctly parsed into a data structure that can be interpreted by another system. For example, social security numbers should contain only numeric data. If the data contains letters, such as xxx-xx-xxxx, the data could be considered to have interpretability problems.
• Recency—The condition measures whether the data is up to date. For example, if an individual moves but the address you have in your system contains the person's old address, the data could be considered to have a recency problem.
• Uncategorized—Choose this option if you do not want to categorize this condition.
• Uniqueness—The condition measures whether there is duplicate data. If the dataflow could not consolidate duplicate data, the records could be considered to be an exception.