Browsing by Author "Sydänlammi, Hertta"

The social urban structure of Helsinki has experienced a significant rise in spatial differences during the last two decades. This development has reflected on schools as rising differences between schools’ student compositions and learning outcomes. Additionally, signs of independent school effects have been observed in several studies. The differentiation of student compositions is feared to exacerbate residential segregation and differentiate schools’ operating environments further. It is possible, however, to intervene this development by drawing the school attendance districts such that the social differences between schools’ student compositions are effectively minimized. For this purpose, new machine learning based optimization tools are needed. The main objective of this master’s thesis study is to examine the possibility to optimize Helsinki’s school districts toward more internally heterogeneous and externally homogeneous social compositions. For this purpose, I have developed an optimization model that minimizes the variance of social variables between school districts by iteratively redrawing the districts’ borders. In a pilot application of the model I optimize the school districts of Helsinki by using the share of population with immigrant background as the optimization variable, while existing school infrastructure (the school locations and student capacities), spatial contiguity of the districts, and school-specific maximum travel distances are used as constraints restricting the shapes that the districts can take. The core finding of this study is that in Helsinki, the social compositions of school districts can be significantly evened out by redrawing the school district borders. However, for the model to be suitable for district planning in practice it needs further development. At this stage, the main limitations of the model are related to the shapes of the optimized districts, the model’s time complexity and the lack of a constraint or optimization parameter that accounts for the safety of children’s school trips.