Capstone Archive

🎓 Welcome to the archive of past capstone projects! Here, you’ll find innovative and impactful work created by GW Data Science students—perfect for inspiration, learning, and celebrating what’s possible through data.

All Semesters Fall 2024Spring 2024Spring 2025 All Subjects DLGISGNNMLTS All Levels GraduateUndergraduate

Deep Learning for Post-Fire Assesment: A CNN Classification of Buildings in the Pacific Palisades

This study aims to segment and classify buildings affected by the 2025 Los Angeles wildfires using deep convolutional neural networks (CNNs).

Spring 2025

GIS

Undergraduate

Author(s): Stella Meserve, Sarah Courtney

Instructor: Sushovan Majhi

Links:

Advancing Temporal Forecasting: A Comparative Analysis of Conventional Paradigms and Deep Learning Architectures on Publicly Accessible Datasets

This study aims to compare the performance of classical time series models (AR, MA, ARMA, ARIMA), modern deep learning techniques (LSTM, Bi-LSTM, Seq2Seq), and a state-of-the-art Transformer model.

Fall 2024

TS

Graduate

Author(s): Liang Gao, Reza Jafari

Instructor: Amir Jafari

Bibliography: Gao, L., Jafari, R. & Jafari, A.H. Advancing temporal forecasting: a comparative analysis of conventional paradigms and deep learning architectures on publicly accessible datasets. Neural Comput & Applic 37, 18173–18184 (2025). https://doi.org/10.1007/s00521-025-11324-0

Links:

Optimizing Fantasy Football Decisions: Performance, Injury Risk, & Hidden Gems

This study aims to develop a data-driven platform for fantasy football enthusiasts that delivers actionable insights, enabling informed lineup decisions, injury risk management, and performance predictions.

Fall 2024

ML

Undergraduate

Author(s): Rohan Rajesh

Instructor: Sushovan Majhi

Links:

Leveraging Graph-based Learning for Credit Card Fraud Detection: A Comparative Study of Classical, Deep Learning and Graph-based Approaches

This study aims to address the problem of credit card fraud detection using classical machine learning, deep learning, and graph neural network (GNN) approaches.

Spring 2024

GNN

Graduate

Author(s): Sunisha Harish, Chirag Lakhanpal

Instructor: Amir Jafari

Bibliography: Harish, S., Lakhanpal, C. & Jafari, A.H. Leveraging graph-based learning for credit card fraud detection: a comparative study of classical, deep learning and graph-based approaches. Neural Comput & Applic 36, 21873–21883 (2024). https://doi.org/10.1007/s00521-024-10397-7

Links:

A Benchmark for Graph-based Dynamic Recommendation Systems

This paper aims to provide a comprehensive benchmarking framework for evaluating the performance of graph-based dynamic recommendation systems.

Spring 2024

GNN

Graduate

Author(s): Tyler Wallett

Instructor: Amir Jafari

Bibliography: Wallett, T., Jafari, A. A benchmark for graph-based dynamic recommendation systems. Neural Comput & Applic 36, 22829–22844 (2024). https://doi.org/10.1007/s00521-024-10425-6

Links:

Multimodal Fusion: Advancing Medical Visual Question-answering

This paper explores the application of Visual Question-Answering (VQA) technology, which combines computer vision and natural language processing (NLP), in the medical domain, specifically for analyzing radiology scans.

Spring 2024

DL

Graduate

Author(s): Anjali Mudgal, Udbhav Kush, Aditya Kumar

Instructor: Amir Jafari

Bibliography: Mudgal, A., Kush, U., Kumar, A. et al. Multimodal fusion: advancing medical visual question-answering. Neural Comput & Applic 36, 20949–20962 (2024). https://doi.org/10.1007/s00521-024-10318-8

Links: