From local landscapes to global systems, Geography students examine how people, place and environment intersect through applied, hands-on work.
In GEOG 490, Senior Research or Field Practicum, students partner with a faculty advisor to complete a culminating experience tailored to their interests. This may take the form of an independent research project, internship or study abroad program. Along the way, they refine their research, analyze data and produce key deliverables.
This experience emphasizes practical application, giving students the opportunity to engage directly with real-world issues across environmental and human geography.
The CISE Showcase, held in April, highlights this work as students share their process and outcomes with faculty, peers and the broader community.
Below is a featured Geography capstone project that demonstrates the depth of inquiry and applied learning within the program.
STUDENT: Savannah Walker
ADVISOR: Dr. Xiaojing Tang
SPONSOR: NASA
White roofs are among the most common ways to mitigate the impact of global warming and urban heat island effects. However, the spatial variability in their cooling effectiveness across cities and within neighborhoods is not well understood.
White roof implementation typically occurs at small scales, such as individual buildings or part of redevelopment projects that do not result in categorical changes in the land cover and land use. This makes detection using remote sensing data challenging.
This project uses time-series Landsat data to identify newly installed white roofs across Boston over the past 12 years. It assesses the cooling effectiveness using thermal remote sensing data from Landsat TIRS.
The research provides valuable information to help city planners better prepare for climate change and achieve their climate goals.
STUDENT: Ethan Dorn
ADVISOR: Dr. Galen Murton
Liminal space is a unique internet phenomenon that has gained popularity in recent years for its ability to evoke a very specific yet unplaceable sensation. Many have attempted to describe this feeling using terms such as the uncanny or kenopsia, but none so far truly encapsulate the paradoxical nature of liminal spaces: comfortable yet unnerving, nostalgic yet utterly unrecognizable.
Liminal space is fascinating for its ability to have established itself as a tangible aesthetic yet, to this day, still eludes a proper definition. This project is a critical examination of liminal space, drawing on concepts and analyses from anthropology and geography.
Informed by the work of Victor Turner on liminality as a social process and Yi-Fu Tuan on space and place, this project examines not only the physical characteristics of liminal spaces but also the social and psychological contexts in which they emerge. It also considers the broader societal conditions that allow and often facilitate liminality to take manifest as a place.
STUDENT: Reagan Kitts
ADVISOR: Zachary Bortolot
Accurately evaluating land classification maps is essential for understanding environmental change, urban growth and resource management. However, many existing tools for accuracy assessment are difficult to use, require specialized software or lack transparency in how accuracy statistics are calculated.
This project addresses that gap by developing an interactive, user-friendly tool for assessing classification accuracy in remote sensing data. The tool allows users to load reference and classified images, generate or manually select sample points and examine underlying data values at each location. By clearly showing how pixel values relate to classification categories, the tool helps users understand how mapped classes are assigned rather than treating them as a “black box.”
The application automatically calculates key accuracy metrics, including overall accuracy, confusion matrices and per-class performance measures - allowing results to be exported for further analysis.
By simplifying and clarifying the accuracy assessment process, this project improves accessibility for students, researchers and practitioners. It supports more informed decision-making in fields such as environmental monitoring and land-use planning, where reliable classification data is critical.
