|
Abstract ID: 77 Abstract Number (within Poster Category): Innovation Application-3 Abstract Title: Course Evaluation Data Visualizer Student: Julia Brand (JABRAND@email.sc.edu) Faculty Mentor: Ronald Erdei (erdei@uscb.edu) Author List: Julia Brand, Gaetano Hirshout, Dr. Ronald Erdei (Faculty Member) Program: Computer Science Abstract Category: Innovation Application Abstract: This project explores how data visualization can shape and enhance student course evaluation results. Legacy survey data is used to develop a dashboard which presents evaluation results over time for individual classes. The visualization organizes historical data into a longitudinal format that supports clearer identification of trends at the course level. Compared with static reports that capture only one section of a course, this proof of concept demonstrates how interactive dashboards showcasing this academic assessment data can improve the utility of the data through enhanced accessibility and interpretability. This project provides a foundation for future development of a production level solution available to Professors at USC Beaufort. |
|
Abstract ID: 22 Abstract Number (within Poster Category): Innovation Application-4 Abstract Title: Wishful Sinking: An Angler's Mental Health Student: Davey Buchanan (DAVEYDB@email.uscb.edu) Faculty Mentor: Brian Canada (bcanada@uscb.edu) Author List: Davey Buchanan, Dr. Brian Canada (Faculty Mentor) Program: Computer Science Abstract Category: Innovation Application Abstract: Wishful Sinking is a fishing simulator, partially inspired by popular indie video games like Stardew Valley, that has a subtle message about mental health and self-improvement. The player controls an angler in a worn-down boat, where catching fish and objects gradually makes repairs and helps the player progress toward making it home. Gameplay involves positioning the boat and using timing-based controls to successfully hook and reel in moving targets. This project was developed in the Java programming using the Greenfoot API as a final project for Dr. Canada's Fall 2025 CSCI/ISAT B145 course (Object-Oriented Programming I). A key technical feature is the use of a Markov chain to control fish behavior. Fish move between three states — wander, patrol, and flee — each with a probability of transitioning to another state over time, making movement patterns feel dynamic and unpredictable, thus making the game more challenging and interesting to the player. The game also incorporates historical and cultural symbols, including Ganesha, the Judaculla Petroglyph, the Sator Square, and the Emerald Tablet. These artifacts are discovered through gameplay and provide upgrades, such as improving speed to assist the player in capturing fish. Finally, the game also supports tangential learning, which in this context means players are introduced to real-world ideas and concepts in a way that feels natural instead of "forced." In Wishful Sinking, this happens both through gameplay and short in-game explanations. Players discover cultural artifacts like the Emerald Tablet or Ganesha while fishing, and each item provides a brief description along with an upgrade. This allows players to learn passively through play while also giving just enough information to spark enough curiosity in the player that they might consider educating themselves further about the subject matter after they have finished playing. |
|
Abstract ID: 18 Abstract Number (within Poster Category): Innovation Application-1 Abstract Title: Training a Domain Specific LLM Student: Cole Jones (COLEWJ@email.uscb.edu) Faculty Mentor: Paul Comitz (pcomitz@uscb.edu) Author List: Michael Maxwell, Matthew Lisko, Cole Jones, Paul Comitz (Faculty Mentor) Program: Computer Science Abstract Category: Innovation Application Abstract: The University of South Carolina Beaufort (USCB) is conducting research on artificial intelligence applications in cybersecurity, led by faculty and a dedicated student team. The project centers on developing a domain-specific large language model (LLM) to enhance cybersecurity education through improved Capture the Flag (CTF) events. Building on successful CTFs from recent semesters, the next event is scheduled for late April 2026. The core feature of this research is the construction of a large language model (LLM) from open source components. The LLM will be fine-tuned using Capture the Flag data, and general cybersecurity data taken from multiple sources. The research team will compile and prepare this corpus of data. A wide variety of sources will be utilized. Key anticipated benefits include: Enhanced offensive security capabilities through exposure to real-world vulnerability patterns, aiding automated penetration testing, vulnerability detection, code analysis, and threat simulation. Significant educational value, equipping students with in-demand AI and cybersecurity skills to strengthen the Lowcountry and national workforce. A bridge to practical real-world tools, improving threat intelligence and defensive simulations. Inspiration for student-led innovation, encouraging new businesses and services in the cybersecurity space. The capabilities above will be used, wherever practical, to design and develop Capture the Flag Cybersecurity events. This project is being performed in the Spring semester of 2026. The project may be continued beyond Spring 2026. |
|
Abstract ID: 51 Abstract Number (within Poster Category): Innovation Application-2 Abstract Title: Project Galileo: A Web-Based Organizational Tool to Search the Stars Student: Michael Maxwell (MM401@email.uscb.edu) Faculty Mentor: Brian Canada (bcanada@uscb.edu) Author List: Michael Maxwell, Daniel Scheer, Nathaniel Gosdin Program: Computer Science Abstract Category: Innovation Application Abstract: For the final group project in Dr. Brian Canada’s CSCI B570 Spring 2026 course (Software System Design and Implementation), we present a work-in-progress for Galileo, a full-stack web application for space enthusiasts who are looking for a convenient online place to schedule, store, and organize their celestial observations. Existing celestial calendars and planning tools provide extensive databases tracking celestial events and objects traversing space. However, many of these services lack information about the visibility of the event from a location on earth. Our Galileo web app provides this key functionality that is accessible through a simple, intuitive interface. The application fetches celestial events from various databases for the user to bookmark, subset, and organize into Galileo’s backend database that is hosted in the cloud. We also developed the Galileo backend API (short for “application program interface”) to provide functionality for approximating event visibility given an event, a time and date, and a set of longitude and latitude coordinates. Our goal is for Galileo to provide space enthusiasts with a highly functional and easy-to-use mechanism to schedule and record their observations of celestial objects. Here we present details about the functionality of our work-in-progress while discussing some of the agile software engineering practices and modern tools that our group used to bring Galileo to life. |