Reliable forecasting models are needed to achieve the climate related goals in the face of increasing transport demand. Such models can predict the long-term behavioural response to policy interventions, including infrastructure investments, and thus provide valuable pre-dictions for decision makers. Contemporary forecasting models are mainly based on national travel surveys. Unfortunately, the response rates of such surveys have steadily declined, implying that the respondents become less representative of the whole population. A particular weakness is that it is likely that respondents with a high valuation of time are less willing to respond to surveys (because they have less time available for such), and therefore there is a high chance that they are underrepresented among the respondents. The valuation of time plays an important role for the cost benefit analyses of public policies including transport investments, and there is no reliable way of controlling for this uneven sampling of time preferences. Fortunately, there is simultaneously an increase in the number of signals sent between mobile phones and network antennae, and research has now reached the point where it is possible to determine not only the travel destination but also the travel mode based on mobile phone network antennae connections. The aim of this thesis is to investigate if and how mobile phone network data can be used to estimate transportation mode choice demand models that can be used for forecasting and planning. Key challenges with using this data source in the context of mode choice models are identified and met. The identified challenges include uncertainty in the choice variable, the difficulty to distinguish car and bus trips, and the lack of information about the trip purpose. In the first paper we propose three possible model formulations and analyse how the uncertainty in the choice outcome variable would play a role in the different model formulations. We also conclude that it is indeed possible to estimate mode choice demand models based on mobile phone network data, with good results in terms of behavioural interpretability and significance. In the second paper we estimate models using a nested logit structure to account for the difficulty in separating bus and car, and a latent class model specification to meet the challenge of having an unknown trip purpose. 

Forecasts of the demand for long-distance travel are a key component enabling the calculation of social costs and benefits of policy actions such as infrastructure investments. Traditionally, such forecasting models have been based on travel survey data. However, response rates to travel surveys have been in decline for decades, calling into question whether the sample of respondents is really representative of the full population. As such, there is a need to explore alternative data sources. One promising alternative is mobile phone network data, which is collected without the need of active participation from the traveller. However, mobile phone network data in this thesis lacks trip and traveller specific information such as trip purpose, socio-economic information, travel party size and mode. Furthermore, it is difficult to distinguish between bus and car trips even at a later stage of data processing, as the two modes share the same infrastructure. 

The objective of this thesis is to investigate the use of mobile phone network data for long-distance mode choice modelling. More specifically, we investigate the specific aspects of mobile phone network data as a source of mode choice travel information in the first research paper of this thesis, how uncertainties connected to the identification of the used mode matter, and how it can be handled in the model. In the second research paper of this thesis, a full-scale Multinomial Logit mode choice model is implemented and evaluated, including the development of how to handle mobile phone network data-specific challenges in the dataset of this thesis, such as the lack of distinction between bus and car trips and the lack of trip purpose information. Once this full-scale mode choice model based only on mobile phone network data has been evaluated, a method for combining mobile phone network data with survey data is proposed in the third research paper of this thesis, and the joint model is compared to the mobile phone network data model in terms of behavioural credibility. Finally, it is investigated whether machine learning can be useful in modelling mode choices using the two data sources in the fourth research paper of this thesis. 

Prognoser för reseefterfrågan av långväga resor är en nyckelkomponent för att möjliggöra beräkningen av samhällsnytta och samhällskostnader i policyåtgärder såsom infrastrukturinvesteringar. Historiskt har sådana prognosmodeller baserats på resvaneundersökningar, men svars-frekvensen i dessa undersökningar har sjunkit under flera decennier, vilket gör att det inte längre är sannolikt att urvalet av svarande är representativt för hela befolkningen. Därmed har det uppstått ett behov av att undersöka tillämpbarheten av alternativa datakällor för modellering av reseefterfrågan. Ett lovande alternativ är mobilnätsdata, som samlas in utan att den resan-de behöver göra ett aktivt val för att delta. Det möjliggör insamling av stora datamängder, men nackdelen är att specifik information om den resande saknas, eftersom personintegriteten måste skyddas för all data som samlas in utan ett aktivt beslut om deltagande. Data som är relevant ur ett transportmodelleringsperspektiv, men som saknas i mobilnätsdatan är till exempel reseärende, resenärens socio-ekonomi, resesällskapsstorlek och färdmedel. Dessutom är det svårt även i efterbearbetning av data att särskilja färdmedlen bil och buss, eftersom de använder samma infrastruktur.

Målsättningen i den här avhandlingen är att undersöka tillämpbarheten av mobilnätsdata för inrikes långväga färdmedelsvalsmodellering. Mer specifikt så utreds först specifika aspekter av mobilnätsdatan som datakälla till en färdmedelsvalsmodell i avhandlingens första artikel, bl a hur osäkerhet i förbearbetningen av identifieringen av färdmedelsvalet spelar in, och hur denna osäkerhet kan hanteras i modellen. Sedan implementeras och utvärderas en fullskalig logitbaserad färdmedelsvalsmodell baserat enbart på mobilnätsdata i avhandlingens andra artikel. För att möjliggöra detta utvecklas metoder för att hantera bristen på särskiljning mellan bil och buss, och bristen på reseärende i mobilnätsdatan. Efter det utvecklas i avhandlingens tredje artikel en metod för att kombinera mobilnätsdata med enkätdata i en logitbaserad färdmedelsvalsmodell. Den samskattade modellen jämförs sedan med modellen som baseras enbart på mobilnätsdata i termer av beteendemässig rimlighet. Slutligen utreds huruvida alternativa modelleringsmetoder i form av maskininlärning kan vara användbara i den variant av färdmedelsvalsmodell som baseras på båda datakällorna.