Please contact Dr. Works (keworks@fsu.eu) for additional help: Submission navigation links for Research Symposium Program Portal WF ‹ Previous submission Next submission › Submission information Submission Number: 117 Submission ID: 8156 Submission UUID: 75624e1e-b91b-4038-a300-64120616a462 Submission URI: /student-research/symposium/research-symposium-program-portal Submission Update: /student-research/symposium/research-symposium-program-portal?token=qwdkb2aK8QkHndZ21uq1r-qFt4Cfd31takYgSUeDbBE Created: Sat, 02/08/2025 - 10:59 PM Completed: Sat, 02/08/2025 - 10:59 PM Changed: Mon, 04/14/2025 - 12:36 PM Remote IP address: 50.4.43.202 Submitted by: Anonymous Language: English Is draft: No Webform: Research Symposium Program Portal WF Submitted to: Student Research Symposium Program Portal Primary Student Contact First Name Primary Student Contact Last Name Pronouns Primary Student Contact FSU Student Email Photo of all individuals presenting this work IMG_3881.jpg115.04 KB Remove Upload requirementsOne file only.2 MB limit. Major(s) of all individuals presenting this work Bio of all individuals presenting this work Hi, I am Robbie Bartels, a graduating senior this year in Computer Science. My research has centered on optimizing database architectures for machine learning applications, working to create more efficient systems for processing and analyzing large datasets. This intersection of structured data management and predictive modeling has become my professional passion, driving me to develop technologies that can scale effectively while maintaining data integrity. Poster Title Abstract This study presents a two-phase approach to analyzing and predicting airline flight delays. The first phase consists of a comparative performance analysis between Neo4j and SQL in tracking cumulative flight delays across aircraft tail numbers. The study measures execution time and query efficiency in calculating cumulative delays across multiple aircraft. This database performance comparison provides insights into the scalability and efficiency of graph-based versus relational database approaches for flight delay tracking. The second phase evaluates three machine learning models for delay prediction: Random Forest, Gradient Boosting, and XGBoost. These models were trained on flight data including basic flight information such as departure times, carrier details, and route distances. The study performed feature engineering to create additional predictors like weekend flight indicators and time-based features. Performance comparison between the three models was conducted using metrics like RMSE and prediction accuracy within various time thresholds. Feature importance analysis across all three models helped identify the most crucial factors in predicting flight delays. Research Mentor Name Research Mentor's College (or High School) Research Mentor's Department (or Subject) Research Mentor's Email Additional Research Mentor(s) Co-presenter(s) Keywords Poster Session/Number Work Complete Exploratory (the research question has been identified and design of approach is outlined) Presentation Modality Face to Face Poster session Synchronous Online Presentation Asynchronous Online Presentation Poster PDF Presentation Template FSU_20241114 1 (2).pdf385.52 KB Remove Upload requirementsOne file only.100 MB limit. Poster Thumbnail Screenshot 2025-03-21 075405.png371.63 KB Remove Upload requirementsOne file only.2 MB limit. I will be printing my poster CAPTCHA What code is in the image? Enter the characters shown in the image. This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Save Leave this field blank
