Akademik Veri Yönetim Sistemi
14th INTERNATIONAL CONGRESS ON ENGINEERING AND TECHNOLOGY MANAGEMENT, İstanbul, Türkiye, 15 - 16 Kasım 2025, cilt.1, sa.1406, ss.221-229, (Tam Metin Bildiri)
The aim of this study is to evaluate the earthquake preparedness level of existing buildings in the city center of Erzincan using modern data processing approaches. Erzincan is located on one of Turkey's most active seismic belts, and assessing the safety of its building stock is of great importance for disaster risk management. For this purpose, two different street survey methods have been integrated into a software system developed within the study. The first method is a questionnaire-based street survey prepared by the ITU Alumni Association Bursa Branch, while the second is the visual street survey developed by Erdem Erdoğan for buildings in Ankara. In addition to these methods, an artificial intelligence-supported module, developed with Python-based machine learning algorithms, was used to analyze photographs of buildings taken from four directions and predict their structural safety. The system was developed using opensource Python libraries, integrating both visual recognition and data classification functions. The analysis results revealed that the predictions made by the AI model were highly consistent (approximately 90%) with the other two methods. These findings suggest that AI-assisted street survey applications can provide faster, more systematic, and reliable evaluations compared to traditional fieldwork. Furthermore, with its user-friendly interface and data analysis modules, the system has the potential to serve as a widely applicable decision support tool in engineering applications. In conclusion, the developed system could significantly contribute to urban-scale building inventory creation, disaster risk management, and urban transformation planning.