Visualizing Complex Origin-Destination Flows: A Network Backbone Approach

Keywords: OD Flow, Network Backbone, Flow Visualization, Commuting, Complex Networks
Abstract Type: Paper Abstract

Abstract

Visualizing Origin-Destination (OD) flows is the first step toward understanding spatial interaction patterns. It is, however, a challenging task, especially for visualizing large amounts of flows corresponding to complex spatial processes. Traditional approaches to visualizing complex OD flows involve several strategies to reduce visual clutter, such as by using a cutoff threshold to filter out small flows or using a better symbology based on line color and width to highlight large flows. These methods, however, may either remove small but significant flows or require intensive steps adjusting symbology parameters, thus largely limiting their applicability. To address such research gaps, this research proposes a network backbone approach in complex networks to the extraction and visualization of the fundamental structure of the flow network for a simplified representation. To illustrate this approach, the disparity filter network backbone extraction algorithm is applied to visualize the commuting flows between census tracts in the state of Florida from the 2012-2016 Census Transportation Planning Products. An important feature of the method is it emphasized the local importance of the nodes and edges rather than just the weight in determining which edge should be kept. The resulting disparity filter backbone commuting network (DFB) is evaluated by comparing it to the backbone commuting network generated by the widely used global thresholding method (GTB) geographically and statistically.