Applications of network segregation measures

Abstract

In this thesis, the main purpose is to measure the political segregation level of main four party voter groups in Turkey. Political segregation is the degree to which political groups live apart. The segregation index is a kind of probability which shows this apartness level. It is crucial to understand the general situation of the counties in terms of the segregation of different party voters who reside in the country, because segregation might have different effects on socioeconomic variables such as crime level, education level and real estate prices. In order to conduct this analysis, main networks for two big cities of Turkey (Istanbul and Ankara) are constructed. The segregation levels of the voters are measured in the neighborhood base analysis. In addition, correlation between the segregation values are measured for the last three elections of Turkey. Finally I check whether the segregation indices are meaningful measures in terms of predicting the real estate prices in Istanbul.