Energy Data Analytics Lab team takes top prize at 2018 Duke Research Computing Symposium with electricity access project

Posted On:

Thursday, Jan 25, 2018 - 2:44 pm

An Energy Data Analytics Lab team won first prize in the poster competition at the 2018 Duke Research Computing Symposium on Jan. 22.

The poster competition—open to all Duke students, faculty, and staff—annually showcases the work of Duke scholars using information technologies to drive innovative research. 

The poster represents a preliminary phase on a 2017-2018 Bass Connections project. Click image to download the poster as a PDF.

The team's winning poster, "Automating Electricity Access Prediction with Satellite Imagery," communicates findings from their efforts to predict the electrification rates of rural villages by applying machine learning techniques to images of lights at night visible from satellite imagery. This study focused on electricity access in Bihar, India.

This preliminary study is part of the team's Bass Connections in Energy & Environment project, Energy Data Analytics Lab: Electricity Access in Developing Countries Based on Aerial Imagery. This year's project builds on the work of several previous Bass Connections and Data+ teams also affiliated with the Energy Data Analytics Lab at Duke, a partnership among the Energy Initiative (the lab's institutional home), the Information Initiative at Duke, and the Social Science Research Institute. 

Awards for the first prize team include four years' use of a beefy virtual computer in the Duke Compute Cluster (~240 GB RAM and 22 CPU cores, valued at $9,300), as well as credits for storage, CPU cores and RAM during 2018, valued at $1,200. The prize will advance work on the next phases of the project, which will include further refinement of the machine learning infrastructure. 

The team includes undergraduates Fangge Deng, Shamikh Hossain, Prithvir Jhaveri, Ashley Meuser, Harshvardhan Sanghi, Joe Squillace, Anuj Thakkar, and Xiaolan You.  Undergraduate team members are pursuing degrees in computer science, environmental science/policy, electrical and computer engineering, mechanical engineering and materials science, and statistical science. Brian Wong, a master's student in environmental management at the Nicholas School of the Environment, provides mentorship and project management. 

Faculty advisors include Kyle Bradbury (Duke University Energy Initiative), Leslie Collins (Pratt School of Engineering), T. Robert Fetter (Nicholas Institute for Environmental Policy Solutions), Marc Jeuland (Sanford School of Public Policy, and Timothy Johnson (Nicholas School of the Environment). 

Are you a Duke student interested in working with the Energy Data Analytics Lab?
Don’t miss your opportunity to apply for a Data+ project team (summer 2018; deadline Feb. 24) or Bass Connections project team (2018-2019, deadline Feb. 16), which will extend the work of the current team.

Are you a Duke PhD student in a field related to the application of energy concepts or a field related to information science?
You might be a candidate for the new Energy Data Analytics PhD Student Fellows program. Learn more. Deadline: Feb. 28, 2018

Is your company or organization a potential partner? The Energy Data Analytics Lab is seeking partners for future projects. We’d welcome a chance to talk with professionals about how a Duke University team could contribute to your upcoming energy data projects. Contact Energy Data Analytics Lab managing director Dr. Kyle Bradbury for more information. 


Location: Zoom (register for link)
Time: 11:00 am
Location: Ahmadieh Family Auditorium (Gross Hall 107)
Time: 9:00 am to 10:00 am