ARM Returns to Brazil for Second Data Science Workshop


ARM Data Center staff transfer knowledge of big data

Group photo during data science workshop in Brazil in February 2020
ARM Data Center staff mingle with workshop attendees for a group photo at a data science workshop in February 2020 in São Paulo, Brazil. Photos are courtesy of Giri Prakash, Oak Ridge National Laboratory.

In October 2017, Giri Prakash, data services manager for the Atmospheric Radiation Measurement (ARM) user facility, and Bhargavi Krishna, an ARM software engineer, went to Brazil to help conduct a two-day workshop about big-data analytics. During the workshop for researchers and data professionals, they focused on analyzing ARM data from the Green Ocean Amazon (GoAmazon2014/15) field campaign in Brazil.

That first workshop at the University of São Paulo was such a success that ARM Data Center staff were invited back for another data science workshop at the same venue in February 2020. Representing ARM were Prakash; Ranjeet Devarakonda, an ARM data scientist and system architect; and Maggie Davis, an ARM data scientist. All three are also colleagues at Oak Ridge National Laboratory (ORNL) in Tennessee.

The 2020 workshop provided a platform to share knowledge and experience between the ARM Data Center and the Brazilian Atmospheric Data Science group, which consists of scientists from São Paulo Research Foundation (FAPESP) and the University of São Paulo.

“Learning from ARM’s experience in managing sensor-based data is extremely useful for our data management effort,” says Prof. Pedro Luiz Pizzigatti Corrêa, the workshop’s organizer. “Similarly, we are sharing our knowledge and experience with ARM regarding complex data integration, visualization, and analysis frameworks, especially big-data tools and technology that we use for analyzing interdomain data sets such as atmospheric parameters and biodiversity data.”

At the workshop, the ARM team taught best-practice data management tools and techniques to 150 Brazilian participants using ARM data. A co-presenter from Johns Hopkins University added expertise about online collaboration for big data in research and education. Brazilian presenters handled several other topics. Attendees of the three-day workshop included field and data scientists, software engineers, and industry representatives.

ARM Lives the Issues

Working on a computer during a data workshop in Brazil
Maggie Davis (left), an ARM data scientist, works directly with a workshop attendee.

The first morning’s keynote addresses included Prakash explaining ARM’s recent advances in data management and analysis, Devarakonda describing open-source-based data tools and systems, and Davis discussing metadata management and data visualization. In afternoon training sessions, ARM staff and Brazilian scientists used Jupyter-based visualization tools to view and analyze ARM GoAmazon data.

Days two and three saw participants break into smaller groups to focus on their particular big-data issues. For example, one scientist explained what he needed to reach his data objectives, and Davis responded with how ARM solved similar issues. This direct, person-to-person knowledge transfer is compelling because ARM has lived—and worked through—the big-data challenges that other science organizations currently face. Such collaborations help cement relationships built through shared field campaigns, such as the GoAmazon partnership between U.S. and Brazilian organizations, that keep valuable scientific data flowing.

Already from this workshop, the group has planned three journal articles related to atmospheric data management capabilities. The Brazilian team that organized the workshop, led by Corrêa, expressed interest in continued professional collaboration for the next three years with the ARM Data Center. And once back at ORNL, Prakash’s team found that 35 new Brazilian workshop attendees had registered with ARM—without any encouragement.