Data visualization and optimization methods for placing entities within urban areas

Abstract

In the first part of this thesis, I present a portfolio of web-based visualizations that illustrate different data-driven ideas about urban environments. These visualizations are intended to provide the user with unique perspectives about cities and the way they function. I detail the conceptualization, data aspects, and implementation of each of these map visualizations. In the second part of this thesis, I describe an interesting optimization problem of placing entities such as trees or shops within a city. The location of these placements needs to conform to certain constraints enforced by spatial distributions of variables such as population, income, travel times, etc. I then present a heuristic-based optimization strategy, that combines some aspects of Gradient-ascent and Simulated Annealing, to address this problem and attempt to generalize this approach to finding the optimal placements of any entity within a given city. I present some initial results of my optimization algorithm and discuss ways in which it can be further improved.