Search for: All records

Traffic forecasting plays an important role in urban planning. Deep learning methods outperform traditional traffic flow forecasting models because of their ability to capture spatiotemporal characteristics of traffic conditions. However, these methods require high-quality historical traffic data, which can be both difficult to acquire and non-comprehensive, making it hard to predict traffic flows at the city scale. To resolve this problem, we implemented a deep learning method, SceneGCN, to forecast traffic speed at the city scale. The model involves two steps: firstly, scene features are extracted from Google Street View (GSV) images for each road segment using pretrained Resnet18 models. Then, the extracted features are entered into a graph convolutional neural network to predict traffic speed at different hours of the day. Our results show that the accuracy of the model can reach up to 86.5% and the Resnet18 model pretrained by Places365 is the best choice to extract scene features for traffic forecasting tasks. Finally, we conclude that the proposed model can predict traffic speed efficiently at the city scale and GSV images have the potential to capture information about human activities. 

Transit-oriented development has been a widely accepted tool among transportation planning practitioners; however, there are concerns about the risk of increasing residential property values leading to gentrification or displacements. Therefore, it is critical to provide precise investigations of the relationships between public transit and gentrification. Although numerous studies have explored this topic, few have discussed these relationships based on detailed measurements of gentrification from a regional perspective. This study aims to fill the research gap by measuring the gentrification subcategories through a hierarchical definition based on data in the New York–Northern New Jersey–Long Island areas and applying the transit desert concept as the measurement of transit services. Through multinomial logistic regression and machine-learning approaches, findings indicate that the rate of transit deserts in economically disadvantaged neighborhoods is higher than the others. In addition, the impacts of transit services are significant in gentrification but insignificant in super-gentrification. These findings can advance the knowledge of the role of the transit service in different gentrification progresses. Based on these findings, policymakers need to be careful when allocating public transit budgets and note the effects of these investments on neighborhoods with different socioeconomic statuses.