Association between Transportation Infrastructure and its Environmental Health Exposures: Developing a Comprehensive Machine Learning Algorithm

Project Details
Author(s) Shima Hamidi, University of Texas at Arlington
Co-Author(s) Mohammad Tayarani, Cornell University; Amir Shahmoradi, University of Texas at Arlington
CTEDD Funding Year 2019 General RFP
Project Status In Progress
UTC Funding $188,442.10
End Date January 30, 2021


Transportation infrastructure plays a vital role in urban life by providing mobility and accessibility for people and goods, but also bring externalities such as air pollution. Research has been devoted to study negative health outcomes and environmental injustice due to transportation externalities. New methods, however, are still required to overcome the data scarcity and improve the accuracy of existing data which are mostly based on estimations rather than actual exposure effects. Furthermore, while different transportation-related factors are commonly considered individually in estimating exposure impacts, new and innovative approaches are critically needed to predict collective exposures to correlated factors in urban environments.
With such a comprehensive and interdisciplinary understanding of the complexity of transportation-related exposures, we also need to move beyond top-down planning approaches, to ones that incorporate the involvement and behavioral reactions of stakeholders to changes in their living environments. This research proposal aims to develop a new exposure modeling based on artificial intelligence. The deep learning algorithm estimates exposure to different aspects of the transportation system by detecting the features in publicly available ...data such as Google Street View and aerial images. The air quality, active transportation infrastructure, and green spaces are a transportation system aspects that this proposal will estimate population exposure to cover a long range of exposure indices. We will implement our modeling framework in two case studies in Dallas Texas and Washington DC with different urban forms and transportation system patterns so we can compare the outcomes under different conditions. The exposure modeling platform then will be used to launch a user interface that will enable the public user to evaluate their exposure to the transportation system. Read more

Subjects#: Active Transportation, Air Quality, Big Data, Infrastructure, Machine Learning