In order for traffic planners to be able to make informed decisions to develop the traffic system, they must have a comprehensive overview of what people's travel patterns look like now and what it has looked like historically. The travel patterns can be used to find types of passenger transport that could be moved to more energy-efficient modes of transport or to model the effects that an investment in infrastructure could have.
Today, traffic planners use traffic models with in-data from surveys of people's travel habits and traffic counts. The disadvantages of these data sources are that it is expensive to conduct surveys and counts, it also provides a very limited number of observations and therefore the models built from these data can only give approximate estimates of people's travel patterns.
The aim of the dissertation is to expand the understanding of what is needed to process large-scale passive data sources such as cellular network data and smart-card data from public transit systems to analyse travel patterns. There are some challenges with this type of data sources. For example, there is a risk that short trips will not be registered in a reliable manner. However, one can improve the results with machine learning methods and in the dissertation it's showed that it can be done even if no training data is available.
– New large-scale passive data sources such as data from the cellular network and smart-card data from public transit systems open new opportunities to observe travel patterns in a way that can provide a much more detailed understanding of the actual travel patterns, says Nils Breyer who recently defended his dissertation in Infra Informatics at the Department Science and Technology (ITN), Division of Communication and Transport Systems (KTS).
Today, traffic planners use traffic models with in-data from surveys of people's travel habits and traffic counts. The disadvantages of these data sources are that it is expensive to conduct surveys and counts, it also provides a very limited number of observations and therefore the models built from these data can only give approximate estimates of people's travel patterns.
The aim of the dissertation is to expand the understanding of what is needed to process large-scale passive data sources such as cellular network data and smart-card data from public transit systems to analyse travel patterns. There are some challenges with this type of data sources. For example, there is a risk that short trips will not be registered in a reliable manner. However, one can improve the results with machine learning methods and in the dissertation it's showed that it can be done even if no training data is available.
– New large-scale passive data sources such as data from the cellular network and smart-card data from public transit systems open new opportunities to observe travel patterns in a way that can provide a much more detailed understanding of the actual travel patterns, says Nils Breyer who recently defended his dissertation in Infra Informatics at the Department Science and Technology (ITN), Division of Communication and Transport Systems (KTS).
