Yahoo! Inc.

Ankur Barua

Larry Weyer

Sanjay Wahi

A massive amount of information is collected when tracking statistics such as web site usage data. To simplify the analysis of such data, it is advantageous to process and segment the web site usage data. One proposed method of segmenting the data is to create a reverse sorted associative container for each different web section of a web site. Page-views, time-spent, sessions, and other web viewing metrics may be used as keys in the reverse sorted associative container and web viewer identifiers are used as values associated with the page-view value keys. After creating the reverse sorted associative containers, each reverse sorted associative container is segmented on a percentile basis.

Methods of processing and segmenting web usage information

Reports, based on aggregation operations on data in a data warehouse, may be efficiently generated. At least two separate aggregation processes are performed on facts of a fact table. Each of a plurality of aggregation processors accesses a separate partition of the facts of the fact table. Each aggregation processor applies the at least two separate aggregation processes are applied to the partition of the facts accessed by that aggregation processor and, for each of the at least two separate aggregation processes, an intermediate result of applying that separate aggregation process by that aggregation processor is saved. For each of the at least two separate aggregation processes, the intermediate results, of applying the aggregation processing for that process by the aggregation processors, are merged.

Growing community of inventors

Sunnyvale, CA, United States of America

Ankur Barua