Assessing Epidemic Spreading Potential with Encounter Network (Short Paper)

Densely populated urban public transportation systems can provide inducive environments for transmitting viruses via close human contact or touching contaminated surfaces. In network analysis, Betweenness Centrality (BC) has been used as the primary metric to measure a node’s communication with others. This research extends from the concept of BC and develops new measures to assess the risk of transmitting disease through public transportation links. Three new concepts are introduced: source Total Betweenness centrality (TBC), target TBC, and Encounter Network. From a network node (source node), the set of shortest paths from that node to all other nodes composes a sub-graph (tree). The source TBC of this node is defined as the sum of BC of all edges of this tree. Similarly, using the shortest path tree consists of the set of the shortest paths from all nodes to the node as the destination, the target TBC of the node is defined as the sum of BC of all edges of this tree. Both TBC can be weighted by edge characteristics such as travel time or trip volume. Another new concept, Encounter Network, is constructed as the intersection between all source-target pairs of the public transportation network. We use the source TBC of a node to evaluate the relative risk of transmitting the disease from that node to other nodes. In contrast, the target TBC of a node can be used to assess the relative risk of being infected by a virus transmitted from other nodes to that node. A preliminary case study is conducted to illustrate the process and results.