Web usage data was analysed and useful knowledge was extracted, previously by employing machine learning methods, particularly clustering techniques. This system is manually constructed; hence it has limited topic coverage. So, I propose a �Web Community Directory� that collects the data of a usage mining process, which is collected at the proxy servers of the central service on the Web. For the construction of the community models, a data mining algorithm, called �Community Directory Miner� is extended and used for Content-Based Personalised Search (CBPS). The Web directory is viewed as a concept hierarchy which is generated by a content-based document clustering method. The construction of Web community directories is seen here as the end result of a usage mining process the data collected at the proxy servers of the central service on the Web