With the "Istanbul Time Machine" project developed within the Marmara University Population and Social Research Institute, the streets and daily life of Istanbul in the past are recreated in a three-dimensional digital environment. Within the scope of the project, in which historians, artificial intelligence experts and data scientists work together, news about fire, accident, theft, trade and daily life in newspapers of the period such as İkdam, Sabah, Tanin and Tercüman-ı Hakikat are converted into digital data with artificial intelligence-supported technologies.
The aim is to position the streets and buildings where the events took place on a three-dimensional city model by matching the information obtained from newspapers with historical maps and archive documents. The three-dimensional visualization of streets and buildings is based entirely on visual and written archival sources and academic literature of the period.
With the project, whose three-dimensional modeling studies started in Eminönü, it is aimed that users can access information about the events that took place in those places in the past, the purposes of use of the buildings and daily life while wandering the streets of Ottoman Istanbul in a virtual environment.
Marmara University Population and Social Research Institute faculty member Dr. Fatma Aladağ said that although there is a lot of information about the major historical transformations of the city in their research on the city history of Istanbul, it has become difficult to answer questions about daily life on a street scale.
Stating that the Istanbul Time Machine project was born out of the search to find answers to these questions, Aladağ said, "For example, if we entered a street in Eminönü in the early 1900s, what would we see? Where were people drinking coffee, where to eat, or how much was the rent of a house or shop? In which neighborhoods and streets did accidents or theft occur more frequently? These are actually the questions that make up the daily life of a city."
Aladağ stated that one of the richest sources where they could find answers to these questions was the newspapers of the period and said: "We can think of newspapers and the information they contain as the daily flow of the city. This information is only scattered among hundreds of thousands of newspaper pages. In Istanbul Time Machine, we reveal this information with artificial intelligence and digital history methods and bring it together with the city by remodeling the Istanbul of the period in three dimensions."
Stating that the main feature that distinguishes the Istanbul Time Machine project from similar works is that it is not limited to recreating the appearance of Istanbul in the past, Aladağ said: "The user will not only wander around the streets of old Istanbul in three dimensions. When he approaches a building or street, he will also be able to access historical information related to that place in the virtual environment. For example, a water problem in a street, the opening of a shop, a rental house or a missing person's notice, theft, fire, entertainment life. In other words, we will be able to read people's complaints and happiness about the city by wandering around Ottoman Istanbul. Therefore, the three-dimensional model was designed for us not only as a visual and virtual scene, but also as an interface to historical data.”
Aladağ explained that they created a research environment by making use of academic literature and visual and written archives and said, "In three-dimensional city history studies, the basic question is often 'How did this city look in the past?' It's happening. We add a second, very important layer to this. 'So what was happening in this city?' When we think like this, of course, a three-dimensional city is no longer just an image to be viewed. "It turns into a research environment in which you can access historical information." he said.
Stating that they focused on the approximately 60-year period between 1880 and 1940 in the textual analysis of the project, Aladağ said, "In three-dimensional modeling, we are currently focusing on a narrower historical section, the early 1900s. We started modeling in Eminönü. This is one of the most active regions of the period in terms of trade, transportation and daily life, and we encounter it intensively in newspapers. As we process the newspaper corpus, what news density is focused on, where is it concentrated, what are the historical relations? "It is taking shape, we will expand the area of our pilot regions." he said.
Noting that they used artificial intelligence in different stages of the project, Aladağ explained that they first focused on the automatic transcription of hundreds of thousands of pages of text printed in Ottoman Turkish. Stating that they have developed their own artificial intelligence-supported text recognition technologies for this purpose, Aladağ continued: "For this, we are now able to automatically transcribe these pages with an artificial intelligence-supported text recognition technology that we have trained ourselves. We have reached an accuracy rate of 93 percent in Ottoman Turkish, especially in printed texts. Now, we are moving to a more difficult stage and actually developing models on genres such as rika and naskh. As their success rate increases, these models we have developed are not only for use in the Istanbul Time Machine, but also for close reading." "We also want to make it accessible so that other researchers can benefit from it."
Stating that after the transcription process, they tried to identify places, people, places and events mentioned in the newspapers with current artificial intelligence methods, Aladağ said, "Finally, we combine all of these with the three-dimensional city model. Of course, we aim to make the connections within a very large world of sources that will take years for the historian to read one by one, more visible."
Aladağ stated that one of the issues that surprised them the most during the studies was that the newspapers of the period included even the smallest details of daily life. Sharing the interesting examples they encountered in newspaper archives, Aladağ said: "For example, in an advertisement published in 1919, a London English teacher announces that he is giving a lesson in Yüksek Kaldırım and guarantees his students to speak English after 30 lessons. He even mentions the price as 30 kuruş in the advertisement. In another news, we see that the phonograph, a new technology of that period, was made available to the public in Istanbul coffeehouses in exchange for money. Newspapers describe this new device in very colorful terms. Or something else." "Our newspapers are talking about a fight that broke out around Galata because a street vendor threw his bagels into the air."
Aladağ said that with Istanbul Time Machine, they aim to reunite these records with the streets and buildings they belong to. Aladağ stated that they attach importance to the resulting infrastructure being a usable resource not only for academics but also for the Ministry of Culture and Tourism, municipalities, museums and other cultural heritage institutions and said, "In this context, we have started working on using the historical heritage of Ottoman Istanbul in education with virtual reality and gamified experience technologies by designing history experience rooms. Thus, we are preparing an infrastructure to put these experience areas into service both in the public sphere and at all education levels. Our long-term goal is to develop this infrastructure not only for Istanbul but for all other cities." he said.
