Metadata or Meta–Analysis

views updated

Metadata or Meta–Analysis

In today's world of data management, it has become increasingly important to master methods of accumulating, controlling, and accessing information electronically. The key to this process is metadata, or data concerning data. Businesses use metadata every day when filing electronically, searching for documents on their intranets, and collecting customer data for analysis. E-retailers and businesses that interact online depend on metadata to run their companies. According to Nigel Ford's 2008 Web-Based Learning Through Educational Informatics, some common types of metadata found today are:

General: Information pertaining to the data as a whole set, including its title, what language it is in, and its tag or description.
Lifecycle: Who has accessed the data, added to or taken from it, and when.
Technical: Refers to what format the data is in, how long it will last, and other IT information.
Rights: Who owns the data, who has written it, and who owns the rights to publish or distribute it?
Relation: Any information that connects one set of data to another, such as a chapter of a book, a summary of a report, or an addendum to a file.
Annotation: Any added information concerning the usefulness of the data to particular operations.
Classification: Where the data belongs within an organization's framework, pertaining to sources and applications—labeling data by names such as a human resources (HR) document, a financial report, or a customer service analysis.

GROUPINGS

Metadata is most often divided into three categories of use. The first is business metadata, or data that conforms to specific business regulations and is tailored for business use. This is a more common, linguistic type of metadata, based on intuitive relations and organized so that users can search for particular sets of data more easily.

The second type of metadata is database metadata, the labels referring to the database itself and how information is organized within it. IT practices, such as source-system mapping and defining objects for spreadsheets, use this type of metadata. Database metadata is also used for security purposes, such as keeping track of when the users last accessed the data and possibly even for what reason.

The third type of metadata is application metadata, an elaborate type of data that explains what other metadata means. For instance, an example of application metadata is a description/explanation of what the monthly customer service report is, and who has access to it.

ANALYSIS AND ACCOUNTING

Beyond helping users to understand data systems and search for specific information, metadata also plays an important role in analysis. Most analysis programs use metadata to collect the required information, organize it into sets, and conduct tests. Thanks to business metadata, data-mining parameters can be set (such as all data collected on the corporate Web sites from the past three months). This information can also be analyzed using metadata, which can include the number of online users who clicked on Web ads, or where the company made most of its revenue within a given time period. The most common barrier businesses find between themselves and this sort of in-depth information analysis is semantics—the need to apply simple language to the purely mathematical or technical metadata.

Other companies find metadata to be especially helpful in accountability practices. Since information concerning who created the data, who has accessed it, and what has been added or taken away from it, is all included in

metadata, mining such information can easily clear up electronic accountability issues. Most metadata made for business transactions includes every part of the process, a record of each step and when it was completed, which can help clear up misunderstandings. Metadata can also be helpful in analysis accountability; it defines data sets, separating them into proper fields and allowing creators to limit what the data can be used for so that none is mistakenly included in a data mining or analysis process.

To develop efficient metadata systems, companies should first accept a clear data scheme including accepted titles and parameters for what types of metadata should be included. Periodic updates using metadata can then be included in reports, giving managers and employees instant access to beneficial information concerning the company's data.