Cheng-Jun Wang

We find that the flow of attention on the Web forms a directed, tree-like structure implying the time- sensitive browsing behavior of users. Using the data of a news sharing website, we construct clickstream networks in which nodes are news stories and edges represent the consecutive clicks between two stories. To identify the flow direction of clickstreams, we define the "flow distance" of nodes (Li), which measures the average number of steps a random walker takes to reach the ith node. It is observed that Li is related with the clicks (Ci) to news stories and the age (Ti) of stories. Putting these three variables together help us understand the rise and decay of news stories from a network perspective. We also find that the studied clickstream networks preserve a stable structure over time, leading to the scaling between users and clicks. The universal scaling behavior is confirmed by the collected 1,000 Web forums. We suggest that the tree-like, stable structure of clickstream networks reveals the time-sensitive preference of users in online browsing. To test our assumption, we discuss three models on individual browsing behavior, and compare the simulation results with empirical data.

With the widespread use of mobile computing devices in contemporary society, our trajectories in the physical space and virtual world are increasingly closely connected. Using the anonymous smartphone data of $1 \times 10^5$ users in a major city of China, we study the interplay between online and offline human behaviors by constructing the mobility network (offline) and the attention network (online). Using the network renormalization technique, we find that they belong to two different classes: the mobility network is small-world, whereas the attention network is fractal. We then divide the city into different areas based on the features of the mobility network discovered under renormalization. Interestingly, this spatial division manifests the location-based online behaviors, for example shopping, dating, and taxi-requesting. Finally, we offer a geometric network model to help us understand the relationship between small-world and fractal networks.

We use clicks as a proxy of collective attention and construct networks to study the temporal dynamics of attention. In particular we collect the browsing records of millions of users on 1000 Web forums in two months. In the constructed networks, nodes are threads and edges represent the switch of users between threads in an hour. The investigated network properties include the number of threads N, the number of users UV, and the number of clicks, PV. We find scaling functions PV∼UV^θ1, $PV∼N^θ3, and UV∼N^θ2, in which the scaling exponents are always greater than 1. This means that (1) the studied networks maintain a self-similar flow structure in time, i.e., large networks are simply the scale-up versions of small networks; and (2) large networks are more “productive”, in the sense that an average user would generate more clicks in the larger systems. We propose a revised version of Zipf’s law to quantify the time-invariant flow structure of attention networks and relate it to the observed scaling properties. We also demonstrate the applied consequences of our research: forum-classification based on scaling properties.