- Mobility Data Science (Dagstuhl Seminar 22021). Mohamed Mokbel, Mahmoud Sakr, Li Xiong, Andreas Züfle, Jussara Almeida, Taylor Anderson, Walid Aref, Gennady Andrienko, Natalia Andrienko, Yang Cao, Sanjay Chawla, Reynold Cheng, Panos Chrysanthis, Xiqi Fei, Gabriel Ghinita, Anita Graser, Dimitrios Gunopulos, Christian Jensen, Joon-Sook Kim, Kyoung-Sook Kim, Peer Kröger, John Krumm, Johannes Lauer, Amr Magdy, Mario Nascimento, Siva Ravada, Matthias Renz, Dimitris Sacharidis, Cyrus Shahabi, Flora Salim, Mohamed Sarwat, Maxime Schoemans, Bettina Speckmann, Egemen Tanin, Yannis Theodoridis, Kristian Torp, Goce Trajcevski, Marc van Kreveld, Carola Wenk, Martin Werner, Raymond Wong, Song Wu, Jianqiu Xu, Moustafa Youssef, Demetris Zeinalipour, Mengxuan Zhang, and Esteban Zimányi. In Dagstuhl Reports, Volume 12, Issue 1, pp. 1-34, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)
Due to the proliferation of handheld GPS enabled devices, spatial and spatio-temporal data is generated, stored, and published by billions of users in a plethora of applications. By mining this data, and thus turning it into actionable information, the McKinsey Global Institute projects a "$600 billion potential annual consumer surplus from using personal location data globally". Multiple communities, in computer science, outside computer science, and in industry, have responded to the pertinent challenges and proposed solutions to individual problems. These communities include moving object databases, mobile data management, spatial data mining, geography, urban planning, transportation, spatial privacy, and spatial epidemiology. Integrating these communities around the common interest of mobility data science is the best chance to achieve impactful end-to-end solutions to real life problems.
The goal of this five-day Dagstuhl Seminar is to create a new research community of mobility data science in which whole is greater than the sum of its parts. It will bring together established leaders as well as promising young researchers from all fields related to mobility data science. Related fields include mobility data acquisition, data quality, data management, data analysis, privacy, and applications. Currently, these research fields are largely working independently from each other solving individual problems, with less focus on the integrated end-to-end solutions. The lack of integration remains a barrier, such that little of these results have come to real world use. Therefore, it is timely to introduce the term Mobility Data Science as a domain that seeks integrated data-to-insights solutions, rather than solutions of point problems.
The goal is to exchange the knowledge of the different communities, align knowledge with the needs of participants from industry, and to discuss the integration vision, opportunities and challenges. Topics of the seminar will include:
- Requirements for Mobility Data Science (applications, challenges, visions, collaborations);
- Data Acquisition and Preparation (datasets, models, data integration, data quality, uncertainty);
- Data Analysis (data mining tasks, result evaluation, decision making, privacy);
- The Mobility Data Science Ecosystem (towards industry-ready tools, requirements, standards);
- Mobility Data Science against Pandemics (contact tracing, simulation, prediction, prevention).
Seminar participants will discuss and collaborate towards the following seminar outcomes:
- A Research Agenda: that defines a new research field of Mobility Data Science
- A Mobility Data Science Curriculum: To train a new generation of data scientists
- Joint Project Initiatives: To foster multi-disciplinary and global collaborations
- A Mobility Data Science Ecosystem: Towards open source tools and systems
- A Collection of Abstracts, Presentations, and Videos: To disseminate mobility data science
To achieve these outcomes, the seminar will be structured in three parts: Meet, Know, and Do. The Meet part (1 Day) will maximize interaction by having participants give 5-minute self-introductions followed by a “speed dating” session where participants will be divided into groups. The Know part (2 Days) will feature presentations and tutorials by experts followed by discussions of visions and challenges. The Do part (2 Days) will have participants leverage ideas and synergies by working in groups on the above seminar outcomes.
This seminar is nicely co-located with the Dagstuhl Seminar "Mobility Data Analysis: from Technical to Ethical" to be held from Sunday, January 9 to Wednesday, January 12, 2022, organized by Bettina Berendt (TU Berlin, DE), Stan Matwin (Dalhousie University – Halifax, CA), and Chiara Renso (ISTI-CNR – Pisa, IT) with which we plan to hold shared sessions and moments of interaction. Participants will have the chance to simultaneously participate in two seminars.
Mobility data is typically available in the form of sequences of location points with time stamps, that are generated by location tracking devices. The use of mobility data has traditionally been linked to transportation industry. Nowadays, with the availability of GPS-equipped mobile devices and other inexpensive location tracking technologies, mobility data is collected and published ubiquitously, leading to large data sets of volunteered geographic information (VGI).
In general, mobility data science is the science of transforming mobility data into (actionable) knowledge. This knowledge is critical towards solutions for traffic management, disease pandemic mitigation, micro-mobility (e.g., shared bikes and scooters), health monitoring, logistics (e.g., delivery services), to mention a few.
Despite the common goal of acquiring, managing, and generating insights from mobility data, the mobility data science community is largely fragmented, developing solutions in silos. It stems from a range of disciplines with expertise in moving object data storage and management, geographic information science, spatiotemporal data mining, ubiquitous computing, computational geometry and more. Furthermore, there is a disconnect in both industry and science between mobility data scientists and domain scientists or end users for which solutions are designed. Therefore, the goal of this Dagstuhl Seminar was to bring together and recognize the mobility data science community as an interdisciplinary research field, strengthen the definition of mobility data science, and together explore challenges and opportunities in the field. The seminar had two objectives: (1) to build a new research community of mobility data science as amalgamation of the several communities who have been looking at mobility data, and (2) to draft a research agenda for mobility data science. This dagstuhl seminar was the first towards these objectives. The consensus of the participants is that more events will be needed in the future to continue the community building effort.
The seminar was held in the week of January 9 -- 14 , 2022. It had 47 participants specialized in different topics: data management, mobility analysis, geography, privacy, urban computing, systems, simulation, indoors, visualization, information integration, and theory. Due to COVID-19, the seminar took place in hybrid mode, with 8 onsite, and 39 remote participants. Despite the challenge of different time zones of the participants, all sessions were attended by at least 37 participants.
The seminar program is given in Figure 1. In the first day, every participant gave a five-minutes self introduction, research interests, and position statement on mobility data science. The rest of the program consisted of panels, and open discussions. To work around the time zone challenge, the seminar activities were centered in the afternoon of Dagstuhl, which was still possible for the Eastern and CHN time zones. The open discussions slots were planned ad-hoc during the seminar. In particular, the slot on Tuesday was used to define what is mobility data science, or more precisely what is the scope of work of this community.
All working group and panel discussions were moderated to converge towards the seminar goals of defining a research agenda and building a community. The results are summarized in this report.
Organization and Panels
To accommodate for the hybrid mode and the time zone differences, we opted to let the participants choose to participate in one of the following three thematic working groups, each having 14-17 members and led by one of the seminar co-organizers:
- Seminar co-organizers: Mohamed Mokbel, Mahmoud Sakr, Li Xiong, Andreas Züfle
- Working Group 1: Mobility Data Acquisition and Privacy
The scope includes the full cycle of obtaining and preparing mobility data for further processing. Examples include innovative ways of data collection, crowdsourcing, simulation, data uncertainty, data cleaning, and data visualization. It also includes innovative ways of ensuring mobile users privacy as a means of encouraging users to share their data. Results of Working Group 1 are found in Section 4
- Working Group 2: Mobility Data Management and Analysis
This includes the full data pipeline from modelling, indexing, query processing/optimization, and data analysis. Existing solutions for mobility data management were discussed and a way forward for a next generation system for mobility data management was conceived. Results of Working Group 2 are found in Section 5.
- Working Group 3 Mobility Data Science Applications
This working group discussed the broader impacts of mobility data science to improve understanding of human behavior, urban sustainability, improving traffic conditions, health, and situational awareness. Specific applications towards these broader impacts including map making, contact tracing, pandemic preparedness, indoor navigation, and marine transportation were discussed. Results of Working Group 3 are found in Section 6.
For each working group a dedicated panel session was organized which was attended by all seminar participants. In addition, two parallel working group sessions were held for discussions and for planning the writing of this report. The working groups presented and further discussed their results with all participants on Wednesday. Four cross-cutting panels discussed the topics of systems, funding opportunities, industry involvement, and curriculum development. All panels started with presentations of panelists as listed below, seven minutes each, where they expressed their positions concerning questions given by the panel moderator. The rest of the panel time opened the discussion to all participants.
- Mobility Data Acquisition and Privacy Panel. Moderator: Li Xiong
Panelists: Gennady (& Natalia) Andrienko, Kyoung-Sook Kim, John Krumm, Cyrus Shahabi
- Mobility Data Management and Analysis Panel. Moderator: Mahmoud Sakr
Panelists: Walid Aref, Panos Chrysanthis, Christian Jensen, Yannis Theodoridis
- Mobility Data Science Applications Panel. Moderator: Andreas Züfle
Panelists: Sanjay Chawla, Flora Salim, Moustafa Youssef, Demetris Zeinalipour
- Systems Panel. Moderator: Mohamed Mokbel
Panelists: Walid Aref, Dimitrios Gunopulos, Cyrus Shahabi, Esteban Zimányi
- Funding Opportunities Panel. Moderator: Andreas Züfle
Panelists: Johannes Lauer, Mario Nascimento, Matthias Renz, Carola Wenk
- Industry Panel. Moderator: Mohamed Mokbel
Panelists: John Krumm, Johannes Lauer, Siva Ravada, Mohamed Sarwat
- Curriculum Development Panel. Moderator: Mahmoud Sakr
Panelists: Anita Graser, Marc van Kreveld, Martin Werner, Esteban Zimányi
- Taylor Anderson (George Mason Univ. - Fairfax, US)
- Amr Magdy (University of California - Riverside, US)
- Mahmoud Sakr (UL - Brussels, BE)
- Flora Salim (RMIT University - Melbourne, AU)
- Maxime Schoemans (Free University of Brussels, BE)
- Bettina Speckmann (TU Eindhoven, NL) [dblp]
- Marc van Kreveld (Utrecht University, NL) [dblp]
- Andreas Züfle (George Mason Univ. - Fairfax, US)
- Jussara Almeida (Federal University of Minas Gerais-Belo Horizonte, BR)
- Gennady Andrienko (Fraunhofer IAIS - Sankt Augustin, DE) [dblp]
- Natalia V. Andrienko (Fraunhofer IAIS - Sankt Augustin, DE) [dblp]
- Walid Aref (Purdue University - West Lafayette, US) [dblp]
- Eric Auquiere (STIB / MIVB - Brussels, BE)
- Yang Cao (Kyoto University, JP) [dblp]
- Sanjay Chawla (QCRI - Doha, QA)
- Reynold Cheng (University of Hong Kong, HK)
- Panos Kypros Chrysanthis (University of Pittsburgh, US) [dblp]
- Xiqi Fei (George Mason Univ. - Fairfax, US)
- Gabriel Ghinita (University of Massachusetts - Boston, US)
- Anita Graser (AIT - Austrian Institute of Technology - Wien, AT)
- Dimitrios Gunopulos (University of Athens, GR) [dblp]
- Joon-Seok Kim (Pacific Northwest National Lab. - Richland, US)
- Kyoung-Sook Kim (AIST - Tokyo Waterfront, JP)
- Peer Kröger (Universität Kiel, DE)
- John Krumm (Microsoft Corporation - Redmond, US) [dblp]
- Johannes Lauer (HERE - Schwalbach am Taunus, DE)
- Mohamed Mokbel (University of Minnesota - Minneapolis, US)
- Mario A. Nascimento (University of Alberta - Edmonton, CA)
- Siva Ravada (Oracle Corp. - Nashua, US)
- Matthias Renz (Universität Kiel, DE)
- Dimitris Sacharidis (UL - Brussels, BE)
- Mohamed Sarwat (Arizona State University - Tempe, US)
- Cyrus Shahabi (USC - Los Angeles, US)
- Egemen Tanin (The University of Melbourne, AU)
- Yannis Theodoridis (University of Piraeus, GR)
- Kristian Torp (Aalborg University, DK)
- Carola Wenk (Tulane University - New Orleans, US) [dblp]
- Martin Werner (TU München - Ottobrunn, DE)
- Song Wu (Free University of Brussels, BE)
- Li Xiong (Emory University - Atlanta, US) [dblp]
- Jianqiu Xu (Nanjing University - Jiangsu, CN)
- Moustafa Youssef (Alexandria University, EG)
- Demetris Zeinalipour (University of Cyprus - Nicosia, CY)
- Esteban Zimanyi (UL - Brussels, BE) [dblp]
- Dimitris Zissis (University of the Aegean - Ermoupolis, GR)
- Machine Learning
- Other Computer Science
- Data science
- contact tracing