Utskrift från Malmö universitets webbplats www.mah.se

New model for traffic planning could make urban travel more efficient


Understanding future travel patterns is a major challenge for transport planners. In a recently published PhD thesis, Banafsheh Hajinasab presents a model that can accurately predict our daily commuting and travel habits.

"My model is largely based on open data that is available online and can be updated automatically," says Hajinasab, a PhD candidate in computer science at Malmö University. Examples of such data could be changes in bus traffic or travel fares — information that is readily available. 

Data-based models are often used to analyse the impact of initiatives such as moves to reduce car traffic or improve public transport. However, in contrast to the system that Hajinasab has come up with, these models require large amounts of data and are usually designed to predict the effects of specific policy or infrastructure investments.

"My model takes individual travel behaviour into account and is not limited to a specific scenario," she says.

Hajinasab’s model uses simulations based on the individual that also take sociodemographic and personal factors into account. These considerations include income, access to a car/bike, environmental awareness, proximity to public transport, travel time, cost and weather conditions. Or, to put it simply, all the things that affect decisions relating to travel.

The study consists of three hypothetical scenarios: the first imagines the existence of a train track between the Swedish towns of Lund and Dalby; in the second, fuel prices are significantly increased; and in the third, public transportation fares are halved. To test the reliability of her model, Hajinasab compared her data with a Swedish travel behaviour survey. She then applied the three different scenarios. The results showed that 72 per cent of the journeys surveyed confirmed Hajinasab’s calculations. Although the accuracy of the model is good, it could be improved. 

“The survey is not necessarily representative of travel behaviour as it’s old and based on interviews with people regarding a specific trip on a certain day. Had there been data from an app that registered all the trips a person takes during a day, that would have been more useful,” she says.

Hajinasab has presented her model on several different travel forums and at conferences.

“I hope that transportation planners will see the value in this open data and individualised approach. Nobody wants to build a new bus stop that doesn’t get used, so the more precise the tool can be, the better,” she adds.

Text: Magnus Jando

Last updated by Maya Acharya