This paper develops a fully-automated workflow for constructing panels of tree canopy from high-resolution multispectral imagery with limited near-infrared (NIR) training data. The workflow utilizes the tree-pixel detection algorithm developed by Yang et al. (2009) and Bosch (2020) on a large set of U.S. urban areas but modifies it by creating automatic ground-truth masks through various visual graphics techniques that leverage modern high-resolution NIR data. Using a subset of cities that represent the different U.S. climate regions, I quantify the effectiveness of the workflow by implementing the algorithm. The comparison shows that my workflow is the option that leads to better results in terms of accuracy, recall, and precision.

We measure face-to-face interactions as the coincidence in space and time of two cellphone users who have previously called each other or who start doing so after the event. We characterize the quantity, quality (network centrality of contacts met), and variety (probability that two interactions are with different contacts) of face-to-face interactions for every individual. We also characterize interactions defined by the different degrees of strength in the call network and differentiate among strong and weak ties. We show how the individual’s experienced density and other features of the urban environment affects these measures of interactions.

This project studies how the improvements in the urban physical space (the park movement) implemented between 1850 and 1950 contributed to the urban mortality transition in the City of New York. Although these landscapes modifications were motivated by incorrect science (such as reducing miasmas), they may have had positive long-term effects.