Home

The Networks Data Science (NDS) group at IMDEA Networks Institute, led by Dr. Marco Fiore, carries out research at the interface of mobile networking and data science, by applying and tailoring tools from artificial intelligence, machine learning, statistical analysis and data mining to the metadata that flows through modern mobile network architectures. Thanks to this methodological approach, we address research problems along two directions. On the one hand, we characterize, model and forecast the complex dynamics of mobile data traffic, and use the derived insights to improve the design and operation of mobile network architectures. On the other hand, we leverage the rich information available in mobile network metadata to solve challenging problems in social sciences with a computational, data-driven approach. In both cases, our research builds on large-scale measurement data collected in the operational systems of major mobile network providers.

The NDS group is active in the following research areas:

Mobile traffic analysis and modelling
We study how mobile services are consumed at large (metropolitan and national) scales, by studying measurement data recorded in production networks. We aim at unveiling the specific dynamics of traffic demands generated by a variety of apps in space and time [CoNEXT’17]. We also work towards identifying complex commonalities and diversities in the way mobile services are used at different times or in different geographical areas [WWW’19].

Network intelligence
We design data-driven algorithms for the effective, anticipatory and automated orchestration of network resources. To this end, we apply artificial intelligence, machine learning, statistical methods and optimization to traffic data collected in large-scale operational mobile networks. Among notable contributions, we assess the limits of current and future technologies, such as resource utilization efficiency in network slicing [MobiCom’18]. We also develop original solutions for network management hinging on deep learning models, such as predictors of the capacity to be allocated to network slices to minimize operation costs [INFOCOM’19, INFOCOM’20], or lightweight approaches to service classification from traffic aggregates [MobiCom’20].

Remote sensing with mobile network metadata
We take advantage of the rich information embedded in data generated within mobile network architectures to solve problems in social sciences. The nearly ubiquitous coverage and high level of detail of individual and aggregate metadata available to operators allow developing data-driven models that largely improve the current state-of-the-art in disciplines like geography, demographics or economics. As representative examples, we employ mobile network metadata to produce cartographies of mixed land use in urban areas [TMC’17, INFOCOM’17], or estimate the dynamics of population density in cities at order-of-minute granularity [TMC’19].

Human mobility analysis from mobile phone data
We analyze individual and mass mobility of people through the lenses of mobile network metadata. Our activities focus on data cleansing, and include characterizing and mitigating the limitations of sparse, irregularly sampled positioning information available to operators [EPJ Data Science’19], drawing laws of the relationship between the location sampling frequency and the quality of captured human trajectories [GLOBECOM’17], and removing noise generated by scattered device associations to base stations [WoWMoM’19].

Projects

The NDS group is currently involved in the following research projects.

DAEMON – Network Intelligence for Adaptive and Self-learning Mobile Networks
Funding entity / scheme: European Commission / H2020 ICT-52-2020
Duration: January 2021 to December 2023

NetSense – Mobile Network Sensing
Funding entity / scheme: Comunidad de Madrid / Attración de Talento Investigador
Duration: March 2020 to February 2024

BANYAN – Big Data Analytics for Radio Access Networks
Funding entity / scheme: European Commission / H2020 MSCA-ITN-2019
Duration: December 2019 to November 2023

Downloads

The NDS group supports the research community by developing open-access tools and datasets.

Dynamic population density datasets
Open datasets of time-varying urban population distributions extracted from mobile network metadata in the cities of Milan, Turin and Rome, in Italy. The dynamic population data was derived from mobile network metadata, and validated against attendance at public events. The work was carried out in collaboration with Ghazaleh Khodabandelou, Vincent Gauthier and Mounim el Yacoubi. Technical details are in the relevant papers [WoWMoM’16, TMC’19].

Madrid highway road traffic datasets
Several traces of weekly road traffic on three highways close to Madrid, Spain, generated using a dedicated microsimulator, which is also available at the webpage above. These are quite rare datasets of realistic traffic in freeways that mimic car in- and out-flows observed via in-situ measurements. The work was carried out in collaboration with Marco Gramaglia, Oscar Trullols-Cruces, and Maria Calderon. Technical details are in the relevant papers [SECON’14, COMCOM’16, VNC’17].

Bologna Ringway road traffic dataset
A one-hour trace of road traffic in the center of Bologna, in Italy, generated using the SUMO microsimulator. The simulation is based on origin-destination matrices developed by the iTetris and Colombo projects funded by the European Commission. The work was carried out in collaboration with Luca Bedogni, Marco Gramaglia, Andrea Vesco, Jerome Haerri, and Francesco Ferrero. Technical details are in the relevant papers [TVT’15].

Köln road traffic dataset
A complete vehicular mobility trace of the city of Cologne, Germany, generated through OpenStreetMap, SUMO and the TAPASCologne traffic demand dataset. The dataset covers 24 hours of road traffic during a typical working day, with over 700,000 individual car trips reproduced. The work was carried out jointly with Sandesh Uppoor. Technical details are in the relevant papers [VNC’11, TMC’14].