Analysis of facebook activity usage through network and human perspectives

Facebook currently is a worldwide-used online social network. This work analyzes usage of Facebook activities through network and human perspectives. We collect patterns of http requests for 20 Facebook activities, construct network-user sessions, and characterize http requests into activities. For each session, we collect two main statistics, which activity appears first and which activity has the most http requests. Then, we calculate the weighted sum to infer activities likely to be chosen by the user. The weight standardized scores indicate that Notification, Like, Chat, Profile & Wall, and Photos are activities that users tend to do first; they also have more network load than the others. Besides analyzing the activities through the network data, we gather additional data through questionnaire responses. The statistics show similar characteristics to those obtained from the network logs. Thus, from both network and human perspectives, Like, Chat, Profile & Wall, Photos, and News Feed & Home are five Facebook activities that users like to do them first and they have higher demands than the others.