Spatial interpolation of traffic counts based on origin–destination centrality

• Network centrality is modified and used to predict traffic counts.
• For the case study 70 observations were sufficient to produce a good model.
• Roadway links with high external to external centrality exhibit higher traffic volumes.

This paper presents a new method to estimate Annual Average Daily Traffic. Often traffic volumes are estimated based on roadway characteristics, such as number of lanes, speed limit, and adjacent land use. However, for many communities, especially small communities, these attributes are uniform across roadway types and therefore unable to adequately explain observed variation in traffic volumes. The new method uses novel explanatory variables that are intrinsically derived through a modified form of centrality, a network analysis metric that quantifies the topological importance of a link in a network. The new approach requires minimal data collection and is easily executed using a geographic information system. The case study showed high quality results (out-of-sample validation R2 = 0.95). The new approach can be used for various activities related to transportation planning and investment decision making.