Data Profiling has recently become recognized as an important new concept in Data Resource Management. Data Profiling is the application of analytical techniques over real data for the purpose of determining its true content, structure, and quality. Data Profiling begins with data and any available meta data descriptions. The meta data is generally incomplete and inaccurate. At the end of data profiling you should have perfectly accurate metadata as well as invaluable information about actual content and data accuracy violations. Effective use of Data Profiling technology will significantly improve the efficiency in implementing projects that migrate data, consolidate data or replicate data for decision support. It also can, and should be, a foundation technology for a data quality assurance program.
This presentation walks through the types of analytics that can be used for each category of Data Profiling. It contrasts the two techniques of Discovery versus Validation and shows where each can be most effective. Also included are several real world examples of where Data Profiling has discovered important facts about a user’s data that had material impact on projects.
Harry Carter from Evoke Software Corporation will guide us through this interesting topic.