The data and code scripts used for the analysis in the paper entitled "The Mobility Oracle: A Framework for Approximating Human Mobility" by Ioanna Gogousou, Manuela Canestrini, Negar Alinaghi, Dimitrios Michail and Ioannis Giannopoulos. (Paper will be submitted for peer review.)

• These authors contributed equally to this work.

It comprises four folders within the zip file:

data: Contains the datasets used for the analysis (raw data and processed data). These datasets are derived from real-world transport network data. This folder includes serialized Python objects (.p file).

code: Python scripts required to perform the analysis.

results: Routing results used in the analysis and discussed in the associated paper.

plots: Visualizations presented in the paper.

Programming Language: Python 3.11 

For reproducibility, read the READ_ME.txt file included in the zip folder.

All data files are licensed under CC BY 4.0, all software is licensed under the MIT License.Abstract

Urban mobility modeling plays a critical role in understanding transport infrastructure and improving its efficiency and sustainability. While existing tools are effective for mobility modeling, they typically require extensive data acquisition, such as surveys, questionnaires, or tracking, as well as domain knowledge for calibration. Instead, we propose the Mobility Oracle, a framework that algorithmically approximates urban mobility by incorporating human preferences in the routing process. The framework builds on open-source data and generates a synthetic dataset for further analysis. It is reproducible, modular, and flexible, making it straightforward to adapt to different contexts. Both the theoretical components and the practical implementation are presented, along with a case study that illustrates the framework’s potential applications. Validation is carried out for Vienna (Austria) and Munich (Germany), comparing our approach against the official city-wide modal splits and a small tracked dataset within one of the cities. The resulting mode shares show a maximum difference of 10% at the city scale and 1.9% for the tracked sample. These results demonstrate that Mobility Oracle can be a useful tool to approximate human mobility. City planners and decision makers can use it to systematically test and evaluate alternative planning scenarios in different urban contexts.