EVENT GRAPHS: ADVANCES AND APPLICATIONS OF SECOND-ORDER TIME-UNFOLDED TEMPORAL NETWORK MODELS

Abstract

Recent advances in data collection and storage have allowed both researchers and industry alike to collect data in real time. Much of this data comes in the form of ‘events’, or timestamped interactions, such as email and social media posts, website clickstreams, or protein–protein interactions. This type of data poses new challenges for modeling, especially if we wish to preserve all temporal features and structure. We highlight several recent approaches in modeling higher-order temporal interaction and bring them together under the umbrella of event graphs. Through examples, we demonstrate how event graphs can be used to understand the higher-order topological-temporal structure of temporal networks and capture properties of the network that are unobservable when considering either a static (or time-aggregated) model. We introduce new algorithms for temporal motif enumeration and provide a novel analysis of the communicability centrality for temporal networks. Furthermore, we show that by modeling a temporal network as an event graph our analysis extends easily to non-dyadic interactions, known as hyper-events.