D4.1 – Mobility Data Analytics and Learning Services

The objective of this deliverable is to present the first version of the “Mobility Data Analytics and Learning Services” developed in EMERALDS. These analytics and learning tools and services are internally known as “emeralds”, as introduced in previous deliverables, such as D2.1. The emeralds releases are provided as software repositories with release notes and instructions on executing the developed methods. This report accompanies the emeralds releases, v1. The key aspects covered in this report include thorough descriptions of the emeralds and how they advance the state of the art, both with respect to the mobility data science field as well as the practical application in the use cases, reflected by the scientific and technical KPIs.

EMERALDS’s vision is to design, develop and create an urban data-oriented Mobility Analytics as a Service (MAaaS) toolset, consisting of the proclaimed EMERALDS services, compiled in a proof-of-concept prototype, capable of exploiting the untapped potential of extreme urban mobility data. The toolset will enable the stakeholders of the urban mobility ecosystem to collect and manage ubiquitous spatiotemporal data of high-volume, high-velocity and of high-variety, analyse them both in online and offline settings, import them to real-time responsive AI/ML algorithms and visualise results in interactive dashboards, whilst implementing privacy preservation techniques at all data modalities and at all levels of a data workflow architecture. The toolset will offer advanced capabilities in data mining of large amounts and varieties of urban mobility data.

The emeralds developed in WP4 “Extreme Scale Mobility Data Analytics & Learning” include T4.1 emeralds that perform state-of-the-art analytics jobs in the Computing Continuum (i.e., at the Edge, the Fog, and the Cloud), T4.2 emeralds that enable active and federated learning over mobility data, and the T4.3 Mobility AI-as-a-Service platform. In total, this deliverable covers eight (8) emeralds, three (3) of which are also already leveraging the MAIaaS platform.

In the process of developing the WP4 emeralds presented in this deliverable, multiple iterations have been performed between technical and use case partners to ensure proper business and data understanding, as necessary for data science developments according to industry standards (CRISP-DM). The key dataset insights are therefore included in the corresponding emeralds sections.