Skip to Main Content
Recognizing and understanding knowledge flow between user interactions in social networks are valuable for sociology, economy, political science, and marketing. In this paper, we present a methodology in order to extract information and discover knowledge from a web traffic log. Our study is based on traffic and login history logs of Kasetsart University's network during a 7-days period from March 1-7, 2011. The summarized HTTP sessions show 39,046 distinct users together with 25,894 IP addresses. We conduct a pattern analysis in six aspects: The Origin of HTTP Requests, Distribution of HTTP Requests at the level of hostname, Time spent communicating online, Overall Traffic Workload Analysis, Facebook Traffic Workload Analysis and Web Access Patterns. The results reveal many interesting patterns and knowledge from raw data.