Cycling Visionaries Awards – Project Details


Bicycle forecast model for Vienna

A bicycle traffic volume forecast model has been create, which is able to predict the hourly traffic volume on eight different points in Vienna. It depends on the weekday, public holidays as well as from seven weather parameters (precipitation, humidity, temperature, wind speed and direction, sunshine and cloud cover). In addition the most influence weather conditions on bicycle traffic are given.

In recent years the government of Austria’s capital, Vienna, implemented different measures to improve the situation for bicyclists, motivating more people to use this functional means of transport. However, for planning new infrastructure, raising awareness for bicycle use and implementing supportive political decisions, reliable bicycle traffic volume forecasts are indispensable – but currently not existing.

In a master thesis a model was developed to find an appropriate solution for this problem. Based on data from permanent counters, a tool has been created to estimate the hourly bicycle traffic volumes with accurate prediction of up to 90%. It calculates the hourly bicycle volumes (between 7:00 and 24:00) in advance, based on analysis of nine (2002-2010) years of data from eight automatic permanent counting stations spread all over the city. This instrument can simplify future planning for bicycle infrastructure, as hourly bicycle traffic volumes can be estimated in advance.

With a cluster analysis based on the location of the counting station and type of day (weekday, weekend, public holiday or school holidays) different types of the daily time-variation curve of bicycle traffic volumes were identified. The results show basically two different types of the daily variation, dependent on whether bicycle traffic is dominated by leisure-time traffic (one peak curve) or weekday traffic (morning and late afternoon peak). With these input information, a multivariate regression has been made to forecast hourly bicycle-traffic volumes, taking into account different weather variables, such as precipitation, temperature, humidity, wind speed and direction, sunshine, cloud cover as well as the season. Additionally, the most influence weather conditions on the traffic volume are shown. As little as 3% of the total bicycle traffic volume was counted when rainfall was detected. The annual periodic trend of the time curve has similarities with the time curve of the average temperature. Consequently, the weather dependency for cyclists in Vienna is very high.

For the practical use of the model developed, few data are required, as for instance data from the weather report, type of day and if the day lies within the public school holiday period or not, to generate reliable volume-data as a basis for future decisions in bicycle issues. The adaption of the model for any other city with similar characteristics as Vienna is possible.

Voting closed. Thank you!
0 votes

David Moosbrugger

Vienna, Austria

Category: Science, Research and Development

Back to project overview