Oracle Enterprise Data Quality products are easy to deploy and use. They offer all the stakeholders of any initiative related to data management the opportunity to improve their data quality.
Ever since there have been databases and applications for them, there have always been accompanying data quality issues. Unfortunately, such issues are often of varying natures and so require individual solutions. The key differences lie with the data type or domain. The most common database domains with respect to data quality are customer data (or, in a more general sense, party data, i.e. suppliers, employees, etc) and product data. Oracle Enterprise Data Quality products account for these differences and provide solutions tailored to each of them.
Profiling current data makes it possible to understand data quality issues and provides a foundation for preparing data quality rules in order to fix and avoid data defects. It enables the user to understand utilized data, pointing out the key areas of data discrepancy, and facilitating analysis of the impact of those issues on business operations, learning from historical analysis, and the defining of business rules by using data directly. This enables organisations to avoid specifying predefined interrelations between data fields and to quickly identify weak points in existing business processes and implementation of technology.
Oracle Enterprise Data Quality software enables business teams to profile large volumes of data from databases, spreadsheets, and flat files. Phrase profiling, i.e. Oracle’s unique method of recognizing text data, allows the user to identify critical information which is hidden in data fields containing unformatted text. Profiling collects information and statistics about data but does not make any changes. Systematic audits detect key quality metrics, missing data, incorrect values, duplicate entries, and inconsistencies.
The results of profiling and audit processes are presented in the form of clear executive dashboards. Using a web browser, employees and directors can monitor and review current data quality against specified criteria. Data quality dashboards enable the quick identification and fixing of issues before they significantly impact business operations. The graphic view shows data quality trends over time, protecting the organization’s investments in data quality by providing the appropriate information to the right people.
Profiling, audits, and executive dashboard offer the appropriate level of information and legibility required to begin local data quality improvement projects or enterprise-wide data management initiatives.
Oracle Enterprise Data Quality provides a rich palette of functions for data transformation and standardization using accessible reference data and simple graphical configuration. In addition to basic data, text, and date field functions, the user can also use functions for contextual data, such as names, addresses, and telephone numbers. The user can also quickly configure, package, share, and deploy new functions containing rules specific to the user’s data and industry without the need to type any code.
Text data is very rarely found in a legible and ordered form. Typical issues include:
All of these issues can be solved using Oracle Enterprise Data Quality products. By using a method for controlling data to quickly mark or describe data, they make possble the manipulation of a single record by parsing it into many structured elements (and records, if needed) and the standardization of results according to predefined rules. Innovative parsing and phrase analysis technology offer a unique tool for searching for hidden information in any text file, and for the creation of standardization rules, transforming that information into structured data.
Oracle Enterprise Data Quality products can be also used to conduct audits with regard to specific business rules, and to transform data on the fly according to those rules, providing a flexible and universal data quality firewall. Moreover, they also provide the ability to view the entire data quality process as a real-time web service. The results of parsing and standardization processes can be displayed in the form of graphic dashboards, providing a comprehensive, accurate and clear view of the entire business.
Matching is a key component of many projects related to data quality. The matching function can support various activities, e.g. de-duplication, prevention of duplication, consolidation, customer data integration (CDI), and master data management (MDM). Oracle Enterprise Data Quality products are equipped with advanced data matching functions enabling the user to identify matching records and to optionally link or merge matched records against trust rules.
A wide variety of flexible, intuitive tools for processing matches and related functions provide solutions for any data issues. All matching rules are completely open and transparent, enabling business analysts to quickly determine why a given matching rule was executed and to rapidly adjust the matching process to specific requirements. Matches can be completely processed by the system or flagged as potential matches and added to the manual analysis queue.
Oracle Enterprise Data Quality also includes a connection module enabling easy access to data from Oracle Siebel CRM. Audit features make possible the executing of data quality rules and flow control in data quality processes. The dashboard feature displays the results of audit processes in a graphical form, while the real-time web service provides the ability to call entire data quality processes as a real-time service.
Oracle Enterprise Data Quality makes possible the automation of a large number of necessary ‘cleaning’ tasks, while certain data rules and validation cannot be executed if data is missing or incorrect. In such cases records can be placed in the ‘case management’ queue for later manual analysis and correction. Naturally, not all data errors require correction; therefore priorities based on business rules can be set for the queue so that only key data is sent for manual analysis.
The Case Management feature is a perfect addition to ‘clean’ automation and is particularly useful with regard to preliminary data validation and correction as part of a system migration or consolidation where defective data may remain undetected until the database start-up, therefore reducing the new system’s performance.
Data quality issues often involve names and addresses. Ensuring there is a correct address format is just the first step; it is much more difficult to check that an address is real. Oracle Enterprise Data Quality offers a solution, using reference information from postal authorities around the world to confirm that addresses are ‘real’. In addition, for verified addresses there is also a function providing geo-codes for mapping applications. The system can parse, standardize, transliterate, and process addresses from more than 240 countries and territories, i.e. essentially everywhere in the world where people live, and handle them in structured or unstructured formats, in any character set.
In the world of data quality product data poses specific challenges. The rules for managing product data are specific to the category of the product in question. For example, data quality rules for resistors are different from rules for capacitors, which are in turn different from the rules for switches, brackets and all other product categories. Each product category has its own vocabulary, terminology, abbreviations, permitted values, and standardization. Furthermore, product information is usually conveyed using non-standard description fields that require recognition and parsing. To complicate matters further, most data quality scenarios concerning product data do not cover single categories, but instead hundreds or thousands of product categories.
To handle such levels of variability Oracle Enterprise Data Quality uses specialized semantic recognition to quickly detect product categories and apply correct rules in context. It can also use the context to identify the meaning of individual words or phrases, and to ‘learn’ new rules and contexts. Following correct recognition, product information can be transformed and standardized using classification, attributes and descriptions generated in any language for further processing in other systems.
Product data also constitutes a particular challenge for matching and merging product records. In this case the product data extension capability can identify and collect key product information, and create a standardized record. It can also identify identical, similar and related records, and merge them against defined trust rules.
This feature can operate in any language and also has a module for connecting to Oracle Product Data Hub, enabling the loading of clean, standardized and de-duplicated product information to MDM hubs.
Data quality and MDM are closely related to each other. MDM data hubs have to be loaded with high-quality, complete, and standardized information. After transforming a portion of data according to applicable data quality rules, such information should be stored as reference data in the data hub. Oracle Enterprise Data Quality products can be used in connection with any MDM solution; however, they are initially integrated with Oracle Customer Hub and Oracle Product Hub.
When transferring data between systems there is a need to ensure the quality, integrity, and general usability of the transferred data. Similarly, integrating data of unknown quality and integrity has a questionable value.
Oracle Enterprise Data Quality can be used with any data integration or ETL system; however, it is initially integrated with Oracle’s flagship product for transferring and transforming data: Oracle Data Integrator. Such integration allows customers to easily and quickly implement Oracle Enterprise Data Quality products during deployment projects for easy and quick deployment as part of a complete data integration solution.