Présentation
**********THE SEMINAR IS CANCELLED********
GLOBAL HEALTH LAB SEMINAR
Wilbert van Panhuis
Assistant Professor, Epidemiology & Biomedical Informatics
Affiliated Faculty, Public Health Dynamics Lab, University of Pittsburgh
“Open data to accelerate the development and testing
of dynamic models for global population health”
Salle Jules Bordet – Bâtiment Metchnikoff
Invited by: Simon Cauchemez (simon.cauchemez@pasteur.fr)
Institut Pasteur – 25, 28, rue du Dr. Roux – 75724 PARIS CEDEX 15
Abstract :
Modeling real-world systems requires observational datasets that are often difficult to get.
No single repository exists that contains all required data and no search algorithms exist
that will “fetch” all required datasets for a model automatically. Modelers spend much of
their time searching and cleaning datasets instead of developing new analytical
approaches. Organizations that collect observational data about populations, viruses, etc.
are often not the same organizations that develop analytical models, requiring multidisciplinary
collaborations and personal connections to access datasets. The lack of openaccess,
standardized datasets is a major bottleneck for innovation in model development
and testing, and a rate limiter of modeling the world’s biological systems. We established
the Project Tycho data repository to reduce the burden of data collection and curation
among modelers and to catalyze model development and innovation. Since the public
release of Project Tycho in 2013, over 4000 users have registered and used Project Tycho
data for new research and technology development, leading to 45 new scientic works by
December 2017. In May 2018, we released version 2 of the Project Tycho repository, with
additional datasets and a more standardized format. Starting in 2019, we will also be
coordinating the MIDAS (Models of Infectious Disease Agent Study) network of infectious
disease modelers, including an open platform for data and software, the MIDAS Digital
Commons, and a wide range of community outreach and training activities, towards
accelerating the development and use of dynamic models for population health.
**********THE SEMINAR IS CANCELLED********