Data Analytics of Library Resources

Library Analytics is used to analyze the huge amount of data that is collected by most colleges and universities when the library website is browsed. It is really useful to assess and evaluate the usage of different types of library e-resource items such as books, journals etc. In this thesis work we have designed and implemented a multi-threaded Library Analytics system specifically for the Ryerson Library. We have analyzed the library usage data to generate different types of reports through a data visualization engine based on requirements from the librarian. Furthermore, we have paid special attention to the task of e-resource item clustering, which could better organize the resources and potentially be used to recommend interesting items to users. We have compared different clustering algorithms and found that association-rule based clustering is 20% more accurate than others and it also identifies the hidden relationships between different browsed articles.