Abstract:User behavioural data provides important cue for deficiencies in the current software design, and points for improvement in future design. Today's web based software services providers could collect different kinds of service usage data, which is a good source of information which can help developers better understand user's behaviours. This paper takes the example of the evolution of an online dictionary services, explores answers for questions such as: what are the key indicators with regard to the quality of software product design and service content, and how to enhance new product success by understandings obtained from user behavioural data analysis. In this paper, we first summarise the possible types of web log data that can be collected so far, Then we introduce a user behaviour model generated for the dictionary domain based on user behavioural data analysis, then we discuss how data collection function shall co-evolve with the online services, and how specific data analysis approaches such as mining algorithms and statistical functions can help generate interesting analysis results of value to design decision makers.

Understanding Requirements for Online Services Based on Users' Behavioural Data Analysis