The World Wide Web and its associated distributed information services provide rich world-wide online information services, where objects are linked together to facilitate interactive access. Users seeking information in the Internet traverse from one object via links to another. It is important to analyze user access patterns which will help improve web pages design by providing efficient access between highly correlated objects, and also assist better marketing decisions by placing advertisements in frequently visited document. We need to study the user access pattern behavior through examining the web access log file, browsing frequency of web pages and the average duration time of visitor. This thesis offers an architecture to store the derived web user access paths in a data warehouse, and facilitate its view maintainability by the use of a metadata. The system will update the user access paths pattern with the data warehouse by the data operation functions in the metadata. Whenever a new user access path occurs, the view maintainability is triggered by a constraint class in the metadata. The data warehouse can be analyzed on the frequent pattern tree of user access paths on the website within a period and duration. The result is online analytical mining path traversal patterns specified by the analyst. Our experimental and performance studies have demonstrated the effectiveness and efficiency with the following contributions: Development of an architecture of online analytical mining (OLAM) using frame model metadata; methodology (stepwise procedure) of implementing OLAM and the resultant cluster of web pages frequently visited by users for marketing use.