Big Data in Ottoman Urban Studies: A Relational Approach to the Archival Data and to Socio-Spatial Analyses of an Early Modern Ottoman City

This paper focuses on two basic archival sources of Ottoman urban history: avârız tax surveys and surety surveys of Edirne in the second half of the seventeenth century. These tax and surety surveys, which are abundant in the Ottoman archives, contain rich information on the residents of Ottoman cities, including names, titles, occupation, gender, religion, property status and numbers of tax units ( avârızhane ). All this information is given on the basis of the mahalle (neighborhood), which provides a practical point of departure for “reading” the Ottoman city. Each register contains approximately 10,000 household heads, with about 10 different attributes listed for each of them. In addition, the data contain some numerical information (e.g., numbers of tax units), although a majority of it is nominal. While some types of nominal data, like religion and gender, comprise only a few possible variants and can thus be analyzed without further classification, others, such as occupation data, contain hundreds, and thus cannot be analyzed without clustering them. In analyzing the data, the distribution of each attribute in the city and its density in an urban space (scaling) can be presented in the form of ratios. One may perform these analyses in the first stage with conventional statistical methods. However, this study attempts for the first time to achieve two further goals: connecting the data types to each other; and highlighting the distinguishing differences among the mahalles using the methods of hierarchical clustering, correspondence analysis, and creating maps by geographic information systems (GIS) applications—none of which is possible with conventional methods. Such an exploration suits both the relational approach I am trying to advocate here—namely, that all elements in the city must be understood in relation to one another—and my effort to lay out the general features of the Ottoman city. This approach will allow us to see how these attributes are spatially distributed based solely on the guidance provided by the big data available in the sources. In this context, I explore the topographical similarities of the mahalles on the one hand, and the socio-economic features and structures of attribute profiles via the scale of their “corresponding distances” on the other. These topographical vicinities and socio-spatial neighbors resemble and do not resemble each other in the city. This paper discusses the processes, challenges and possible contributions of the application of big data to urban historical studies.