Bike flows and performance of bike facilities:implementation of a procedure for bike traffic counting in the city of Bologna

The current study investigates how employing traffic counters at best to count bike flows. In particular, this paper illustrates the main characteristics of traffic counters, discussing their advantages and shortfalls, and where they have to be located in order to minimize vehicular and pedestrian interferences when counting bike flows. Instruments were placed in 20 sections, for 7 days. Data were collected continuously. The choice of the particular type of counter, either radar or pneumatic, was performed by taking into account both geometrical and operational variables such as bicycle lane width, type of transport mode allowed, and proximity to roads with high traffic volumes. In addition to instrumental bike flow monitoring, a manual monitoring was conducted in each section. Operators’ reports indicate the instant in which bike cross the monitored section, the direction, and the number of cyclists riding outside the bicycle paths. Manual monitoring is useful to compare data, to determine the level of use of cycle tracks, to calibrate instruments and evaluate their precision. Instrumental data consist of numerical strings containing date, time, direction, speed and number of reflection points, which indicate vehicle size. During the instrumental monitoring phase, a first calibration was performed, determining a range of reflection points corresponding to each class of vehicle. In this way we could distinguish between pedestrian, bicycles, scooters, and cars. Secondly, instrumental data were compared with manual data referred to the same time interval in order to improve the results obtained in the first calibration step. Indeed, under the assumption that the error associated to manual data is lower with respect to the instrumental data one, the number of events recorded by instruments is similar to the number of events recorded by operators. In our case, the error regarding instrumental counts was about 10%. For the specific case of radar traffic counters, the number of reflection points changes according to the direction of travel, due to the value of tilt angle, measured with reference to roadway axis. Then, the collected data were introduced in templates, that allowed drawing bar-graphs with the number of bikes every 15-minute interval, during every day of measurement. By processing these data, we obtained daily flows, weekday and holiday average flows, medium frequency, peak hour factor, level of use, peak periods, climatic dependence and other information. The methodology applied in this research is easy to use and can be adapted to the specific goal of the study that will be performed in order to individuate critical situations and possible interventions.