The NetMob 2023 Data Challenge is a competition colocated with NetMob 2023, that will take place on October 4-6, 2023 in Madrid, Spain. The challenge aims at deriving new knowledge from the analysis, characterization, modelling and cross-correlation of an original high-resolution dataset describing mobile service usage in multiple regions of France.

CoCo5G is supporting this event, with Orange and IMDEA Networks directly involved in the organization.

Challenge format:

  • Prospective participants shall apply to the challenge in order to get access to the dataset. Instructions are available at the challenge webpage.
  • Admitted participants shall submit two-page extended abstracts in PDF format including title, author(s), affiliation(s) and email address(es) on the first page, describing their preliminary results from the dataset analysis by June 30, 2023.
  • Selected participants will be invited to submit a final report with complete results by September 22, 2023, to present them during the NetMob 2023 conference, and will compete for the final challenge awards’.

Topics:

Proposals should address one of the following three topics:

  • Understanding social challenges through the lenses of mobile service usage.
  • Generating reliable datasets of synthetic mobile data traffic.
  • Automating network management with data-drive intelligence.

Dataset description

  • 20 urban areas in France
  • 68 mobile services
  • 77 days continuous days
  • 100 x 100 m2 spatial resolution
  • 15 minute temporal resolution
  • 400+ billion data points
  • 2.3+ TB of data

A full description of the dataset is available at https://arxiv.org/abs/2305.06933. Examples of how to manipulate the dataset are available at GitHub.

Important dates:

  • Challenge launch: May 12, 2023
  • Abstract submission deadline: June 30, 2023
  • Notification of acceptance: July 21, 2023
  • Registration deadline: July 31, 2023
  • Final report deadline: September 22, 2023
  • Presentation in the conference: October 6, 2023
NetMob 2023 Data Challenge