Automatic counters (e.g., loop detectors) used for the continuous collection of cycling count data are subject to periodic malfunctions, leading to sporadic data gaps. This problem could affect the calculated values of the annual average daily bicycle (AADB) volumes and impact the estimates of the daily and monthly adjustment factors at these count stations. The impacts become more significant when the data gaps take place frequently and/or for long periods. This research addresses the problem of missing cycling traffic volumes at the count stations that experience frequent sensor malfunctions. The main hypothesis is that a strong correlation may exist among the cycling volumes of nearby facilities within a network. This correlation can be used to develop neighborhood models based on the available historical data. This study made use of a data set of more than 14,000 daily bicycle volumes from the city of Vancouver, Canada. The data were collected between 2009 and 2011 at 22 different count stations. A correlation analysis was first undertaken, and the results showed a strong correlation between the cycling volumes at most of the analyzed facilities. Furthermore, a cross-correlation analysis showed that the strongest correlation between each pair of count stations took place at a time lag of zero days (i.e., concurrent data). Accordingly, a correlation threshold was selected and used to define a set of neighbors for each cycling facility. Statistical models were developed to relate the daily cycling volumes of each pair of neighbors. The models were validated; the mean absolute percentage error (MAPE) was used as an evaluation measure. In general, the MAPE was less than 20% for most facilities when a correlation threshold of 0.6 was used to identify neighbors. However, the error dropped to approximately 15% when higher thresholds were used. The concept should prove useful in estimating the missing cycling volumes in a monitoring program or a data clearinghouse implementation.
Estimation of Daily Bicycle Traffic Volumes Using Spatiotemporal Relationships
2017-09-09
Article (Journal)
Electronic Resource
Unknown
Estimation of Daily Bicycle Traffic Volumes Using Spatiotemporal Relationships
Online Contents | 2017
|Estimation of daily bicycle traffic using machine and deep learning techniques
Online Contents | 2022
|Estimation of Annual Average Daily Bicycle Traffic with Adjustment Factors
Transportation Research Record | 2014
|Development of Daily Adjustment Factors for Bicycle Traffic
Online Contents | 2013
|