Data-facing work in XSEDE Extended Collaborative Support Services (ECSS), Sergiu Sanielevici (PSC)
Description: The Extended Collaborative Support Service (ECSS) improves the productivity of the Extreme Science and Engineering Discovery Environment (XSEDE) user community through collaborations to optimize their applications, their work and data flows, and engages practitioners of disciplines that have not traditionally used advanced cyberinfrastructure (ACI). Novel & Innovative Projects (NIP) has the primary responsibility within ECSS for this latter task. NIP provides mentoring to help projects be successful and advice on the use of technologies such as virtual environments, machine learning, virtualization and containers. INIP is now focused on helping AI and “big data” projects on novel SP resources scheduled to enter production in 2020, including Bridges-2 at PSC and Expanse at SDSC.
Data librarians may become “data science” librarians by necessity, if they are not already serving in this role. In this call, we will provide an overview of the particulars of two libraries’ data science support beyond data science (e.g., across departments, collaborating with IT, etc.) and have an open discussion of how other groups are dealing with similar support needs.
JANUARY 7 – Identify/prioritize topics for the coming year!
Sharing on behalf of the PEARC20 Program Committee:
Deadlines: (Updated Jan 8)
January 22nd: Tutorial submissions due
January 22nd: Workshop submissions due
February 17th: Technical track full paper submissions due
February 17th: Lightning Talk Abstracts submissions due
February 24th: Student technical track full paper submissions due
April 24th: Poster submissions due
April 24th: Student posters submissions due
May 1st: Panel submissions due
May 1st: BOF submissions due
May 1st: Viz Showcase submissions due
May 15th: All camera-ready submissions due
PEARC20 will explore the current Practice and Experience in Advanced Research Computing, including modeling, simulation, and data-intensive computing. PEARC20 will be in Portland, OR from July 26th-30th, 2020. This year’s theme, “Catch the Wave,” embodies the spirit of the community’s drive to stay in front of the new waves in technology, analytics, data, visualization, and a globally connected and diverse workforce.
PEARC20 brings together community thought leaders, CI professionals, and students to learn, share ideas, and craft the infrastructure of the future. The PEARC20 student program will provide students with a range of opportunities to participate in both student activities and the full technical program so that they may share their research efforts and gain insights and inspiration from like-minded individuals at the conference.
The session will include a series of lightning round style case studies from a variety of institutions whose libraries and IT departments are collaborating. This will be followed by a discussion about these approaches and collaboration more broadly.
Bart and Yoh will be talking about the different stages of the research life cycle, mapping the stages to different groups at their institutions, and helping us do a similar exercise for our own institutions. We’ll be discussing how support for research data comes in at multiple stages in the research life cycle and which organizations at our institutions provide that support.
“Data pipeline case studies” with Tylar Murray of USF and Ryan Chard of Argonne National Laboratory. Tylar will give an overview of an Apache Airflow-based pipeline for processing satellite imagery and discuss possibilities for future improvements. Ryan will preview a data automation platform that is currently in alpha release but will eventually be available via a Globus subscription.
CaRCC developed an initial set of committees in October 2017 to focus on four areas as prioritized by a survey that reached a fairly broad audience of research computing and CI professionals, and these are listed below. If you are involved with CaRCC and interested in helping out on one or more committees, let us know at email@example.com!