The use of mobile phone data to characterise the mobility patterns: challenges and limits

Reducing carbon emissions in the transport sector is essential for the EU to reach its climate targets. EU policies focus on promoting sustainable mobility while ensuring that regions across Europe remain well-connected. To effectively shape these sustainable transport systems, it is essential to understand the mobility patterns and the motivations behind them. Travel behaviour – which includes choices about when, where, and how individuals travel – directly influences the flow of people and goods, and is shaped by factors like transport alternatives, personal preferences, and geographic location. The rapid evolution of data collection methods has opened new possibilities for analysing and modelling these behaviours, offering deeper insights to support efficient and sustainable transport systems. In this context, this research aims to examine and compare the potential of two different data collection methods for understanding travel behaviour. Specifically, a survey-based approach and mobile phone data analytics are compared, focusing on Vodafone's mobile network data and the Audimob Observatory survey. Those databases are provided by Ferrovie dello Stato Italiane (FSI), the national Railways, which is the primary partner of this research. The thesis followed a methodology based on four steps. The first phase provides a comprehensive literature review that examines existing research on methods used to analyse travel behaviour, focusing on mobile phone data and traditional survey techniques. In the second step, Tableau software was used to analyse travel data in both Vodafone and Audimob datasets to identify mobility patterns. This phase explored which aspects of travel behaviour each dataset was able to capture. After using Tableau, the third step compared the two approaches to investigate and understand their unique features; to this end, a SWOT analysis was conducted to evaluate the strengths, weaknesses, opportunities, and threats of both methods considered. Finally, the benefits of each approach were assessed to determine their suitability for different research objectives and their value for planning purposes. This comparison ultimately offers insights into the advantages and disadvantages of big data versus traditional surveys in understanding travel behaviour. The results of comparison of the two datasets show how each of them has unique features but using them individually is to properly describe the mobility patterns. Big data provides a large volume of information, which is particularly useful for analysing trips across various times and zones. However, it lacks the precision needed to identify specific details – such as the mode of transport – which can be more accurately captured using datasets like Audimob. Another issue is that despite the valuable data provided by the data sources (Vodafone), there is still a lack of transparency regarding the algorithms and methods used to generate it. This uncertainty makes it challenging to fully understand and interpret the dataset, thereby limiting the depth and reliability of the analysis. Additionally, when comparing the two datasets over the same time, significant discrepancies are often observed, further complicating the reliability of the information provided by big data alone. Therefore, developing a method that integrates both datasets is crucial. By combining them, the complementary strengths of each can be fully leveraged, resulting in more comprehensive and accurate insights.