Bass Connections

Duke students’ Bass Connections research on energy access and data analytics comes together in a final energy presentation on synthetic imagery used to improve automated wind turbine detection in satellite imagery, especially when applied to diverse locations.

Efforts to ensure energy access across the globe are often hampered by a lack of critical information to guide decision-making and electricity system planning. Information on village-level electricity access and reliability, as well as the location and characteristics of power system infrastructure, is especially scarce. Decision-makers require this information to determine the optimal strategies for deploying energy resources, like where to prioritize development and whether electrification should be accomplished through grid expansion, micro-grids, or distributed generation.

During the 2020-2021 school year, a Bass Connections research team at Duke University aimed to develop deep learning techniques that can automatically and rapidly scan massive volumes of remotely sensed data, such as satellite imagery, to develop detailed maps of energy infrastructure. These deep learning approaches may provide powerful tools for researchers, policy-makers, and governments to collect energy systems information. This video captures the Bass Connections team’s end-of-year presentation in April 2021.

The team used machine learning to create a model that detected wind turbines solely from satellite imagery by training it first with real images of turbines. Since these images are scarce and in practice the machine learning techniques need to be applied to different locations than from where the training data are available, this approach was compared to data resulting from a model which also was trained on synthetic images of wind turbines. Synthetic images, while they might look real to the machine, are generated images and are not genuine photos. Feeding the model synthetic images of wind turbines increased the accuracy or “average precision” of the predicted turbine location.

Bass Connections is a unique Duke University program that brings together faculty, postdocs, graduate students, undergraduates, and external partners to tackle complex societal challenges in interdisciplinary research teams.

Student Team Members: Ada Ye (T'23), Jessie Ou (T'22), Wendy Zhang (T'21), Eddy Lin (T'22), Tyler Feldman (T'23), and Jose Moscoso (MIDS '21)

Faculty Team Leaders: Kyle Bradbury (Pratt School of Engineering and Managing Director of the Energy Data Analytics Lab at the Duke University Energy Initiative) and Jordan Malof (Pratt School of Engineering)

Learn more about the project:

• Bass Connections
• The website created by the team

Nearly a billion people in the world lack access to electricity. This global challenge is made all the more daunting by gaps in critical data about existing energy infrastructure. Policymakers and businesses need these data in order to make sound decisions (e.g., whether to expand the national grid, build a microgrid, provide direct off-grid solar PV, or pursue some other solution). 

In this April 2020 webinar, an interdisciplinary team of Duke University students and faculty describe their efforts to use satellite imagery and cutting-edge deep learning techniques to help fill some of these gaps. Their project, “A Wider Lens on Energy: Adapting Deep Learning Techniques to Inform Energy Access Decisions,” is part of Duke’s unique Bass Connections program. Faculty affiliated with Duke’s Energy Data Analytics Lab led the team.

The team built on previous research projects conducted by students and faculty at Duke. Ultimately, with this series of projects, the Energy Data Analytics Lab seeks to build a tool researchers can use to identify and map out worldwide energy infrastructure and electrification needs to supplement ground truth information and support improved decision-making.

The 2019-2020 team sought to improve the existing model’s ability to adapt to different geographies, a goal that requires training the model with highly diverse datasets. Toward this end, team members worked to develop synthetic imagery that would help make the training datasets more representative of a broader range of geographies. In particular, they sought to train the model to recognize a more diverse array of rooftop textures (which differ by location).

Check out their presentation to learn more about their project’s aims, process, and outputs! 

Student team members:

• Ayooluwa Balogun - Mechanical Engineering (E’21) (Participated Fall 2019)
• Aneesh Gupta - Computer Science (T’22)
• Scott Heng - Computer Science & Statistical Science (T’21)
• Vivek Sahukar - Masters Program in Interdisciplinary Data Science (MIDS’20)
• Norah Tan - Computer Science & Mathematics (T’22)
• Gaurav Uppal - Mechanical Engineering (E’20)
• Jason Wang - Computer Science (T’21)
• Winston Yau - Public Policy & Physics (T’22)

Duke faculty team leaders:

• Dr. Kyle Bradbury – Duke University Energy Initiative, Pratt School of Engineering
• Dr. Leslie Collins – Pratt School of Engineering
• Dr. Jordan Malof – Pratt School of Engineering 

Visit the 2019-2020 team’s project website. 
View the team’s data on GitHub.
Learn more about Bass Connections at Duke. 
Learn more about the Energy Data Analytics Lab. 
Join the Duke University Energy Initiative email list for updates on energy news, events, and opportunities at Duke. 

Story+ is a 6-week paid summer research experience for Duke students—undergraduates and graduates—interested in exploring humanities research approaches (archival research, oral histories, narrative analysis, visual analysis, and more). The program combines research with an emphasis on storytelling for different public audiences. In Story+, students are organized into small project teams and have the opportunity to participate in a flexible mini “curriculum” on research methods and storytelling strategies. Want more information? See the Story+ booth at the Bass Connections Fair on January 24, 2020. 

Applications open January 24, 2020 and are due by February 14, 2020. Applications will be evaluated on a rolling basis, so students should apply ASAP. 

This year, two of the Story+ projects focus on energy topics:

Body Work: Reanimating Policy Responses to Coal Mining Disasters

During this collision of artistic and academic energies, students will examine U.S. policy responses to significant coal mining disasters during the 20th Century and experiment with methods of processing their research through dance. Drawing on evidence such as transcripts of Congressional hearings, federal reports explaining the causes of disasters, and oral histories with coal miners and their families, students will employ content analysis methods to answer two primary questions: how were the narratives used to explain each disaster constructed? And how did those narratives influence policy that aimed to prevent similar catastrophes in the future?

At the same time, dance artist, educator, and researcher Justin Tornow will introduce the students to embodiment methods, which will include an introduction to somatic practices, structured improvisations for movement and spatial orientation, and the use of chance operations. By the end of the six-week term, students will draw on these tools to compose a post-modern movement performance that communicates both their research and the results of including embodiment as one of their methodological cornerstones. Through this unique research experience, students will investigate themes such as the politics of expertise, the role of focusing events and class and gender-based power dynamics in policymaking, the impact of embodiment on academic inquiry and communication, and the alienation of human bodies from processes of energy production in fossil-fueled societies like the modern U.S.
Project Sponsors: Dr. Jonathon Free (Duke University Energy Initiative and Justin Tornow (Dance Department)

Joining the electric circus: rural electrification and gender in the papers of Louisan Mamer

Between 1939 and 1941, representatives from the Rural Electrification Agency organized a carnivalesque roadshow designed to encourage families to purchase and use electrical appliances and other equipment in their homes and on their farms. A key audience of the roadshow was rural farm women, who were seen as equal partners in the effort of electrification -- and who, the REA reasoned, needed to be shown the way to modernity through electricity. This Story+ project will draw on the Louisan E. Mamer Rural Electrification Administration Papers located at the Smithsonian National Museum for American History to examine how officials’ understanding of the gendered division of labor on American farms informed the tactics they used to encourage utilization of electricity. The overall goal of the project is to understand and share how assumptions about gendered labor influenced the electric circus’s programming, as well as collate any lessons learned for similar programs happening today.

Students will be asked to (at minimum) compile a report on their findings for the Duke University Energy Access Project, and there is also scope to create a podcast episode, or a brief documentary-style video. There may also be an opportunity to contribute to a collaboration with the Smithsonian National Museum of American History in Washington, D.C.. The Data+ project entitled, “Taking electrification on the road: Exploring the impact of the Electric Farm Equipment roadshow (1939-1941),” is a partner project to this one and may offer opportunity for collaboration with a data-driven team.
Project Sponsors: Dr. Victoria Plutshack (Energy Access Project), Dr. Rob Fetter (Energy Access Project), Dr. Ashley Rose Young (Smithsonian National Museum of American History)

Curious about previous energy-related Story+ projects?

• Coal & America: Stories fro the Central Appalachian Coalfields (Summer 2018)