Istanbul A Way Out

"Istanbul A Way Out" is a research-driven architectural project developed by a team of young architects in response to the 2023 Turkey earthquake and the growing risk of a major seismic event in Istanbul. Instead of focusing solely on structural collapse, the project explores how everyday urban elements—like narrow streets, parked cars, and architectural projections—can block escape routes and delay emergency aid. Using AI-based analysis, these elements are mapped and scored across different neighborhoods, revealing how urban inaccessibility often overlaps with socio-economic disparities. The project offers a new lens for disaster planning—centered on movement, accessibility, and survival in a city shaped by unplanned growth and inequality.



Istanbul, a megacity larger than any European capital, is shaped by layered socio-economic contrasts and decades of unplanned growth. To train our AI model, we selected seven case study areas across the city to capture a wide variety of urban elements. Drawing on local knowledge from team members familiar with Istanbul’s diverse fabric, we began questioning whether our model’s outputs would reflect these socio-economic differences.

To test this, we focused on three contrasting neighborhoods:
Dolapdere, densely populated and underdeveloped;
Ortaköy, historically rich with moderate planning;
and Göztepe, a newly planned, low-density, high-income area.

By embedding this social context into the analysis, we aimed to see if urban inaccessibility after disasters correlates with broader patterns of inequality.



In this Illustruation/Google Maps Image, YOLOv model have identified elements that could block emergency access—like bay windows, narrow pavements, parked cars, or even lamp posts. After training 300 Google Images, Object Detection starts to see the invisible barriers that might slow down help when every second matters .



This graph illustrates the relationship between the physical width of a street and its associated risk score. Narrow and densely occupied streets correspond to higher risk levels, while wider, more open roads are considered safer in terms of emergency accessibility.

This finding reinforces the core argument of A Way Out: in disaster scenarios, the accessibility of streets can be just as critical as the structural integrity of buildings. The project emphasizes that post-earthquake survival depends not only on what collapses, but also on whether rescue teams can reach those in need.




As a result we had risk analysis maps for the neighborhoods of Dolapdere, Ortaköy, and Göztepe. Each area was evaluated using grid systems of varying sizes, and risk levels were calculated using a “Total Risk Score” formula. At the bottom of each map, the formula is detailed: the risk score is normalized based on the proportion of obstructed street, sidewalk, and open space areas within the total image area.

This methodology enables a comparative understanding of neighborhood-level accessibility vulnerabilities, highlighting how spatial configurations impact emergency response potential. By applying this scoring system across different urban textures, A Way Out aims to create a scalable and data-driven tool for post-disaster risk assessment and 
planning.


TEAM

Eren Sezer
MSc Architect / Researcher
Turkey
eren.sezer@hotmail.com



Nour Fneich
MSc Architect / Researcher
Lebanon
nourfneish1@gmail.com



Andrei Calin Teodorescu
MSc Architect / Researcher
Romania
andreiteodorescu1997@gmail.com


Egemen Sezer
MSc Architect / Researcher
Turkey
egemensezeruni@gmail.com


Raşit Eren Cangür
Phd Architect / Researcher
Turkey
rasiterencangur@gmail.com


Sonya Ragimova
MSc Architect / Researcher
Russia
arch.sona.ragimova@gmail.com


Nicolo Carlini
MSc Architect / Researcher
Italy
carlininicolo@gmail.com