Building and evaluating a logging feature

Abstract

Information is an integral part of today's society. People can easily look up information through search systems such as Google, Netflix and Facebook. User need to be satisfied by providing relevant content, features and mechanisms. It is therefore essential that we understand user behavior in search systems. This thesis studies a search system named BIBSURF. BIBSURF is a bibliographic library and was developed to study the possibilities and effects of different information models for bibliographic information, and to facilitate this research a logging feature is needed track how users interact with the system. To design a logging feature, we need to consider what data should be collected, how we can extract meaningful information from the data and other concerns such as privacy. A logging feature was developed with the help of some JavaScript libraries. This logging system was evaluated with a usability test. In addition, a screen recorder software was used to record participants' interactions to compare this with the logger. Several aspects of the logger was analyzed and strengths and weaknesses were identified. This study can give other researchers insight in designing and evaluate loggers to conduct user studies.