UVDAT: Urban Visualization and Data Analysis Toolkit

Urban areas, particularly those with underserved populations and critical assets in vulnerable zones, are at increased risks of cascading failures due to climate change, urbanization, and aging infrastructure. There is an urgent need for urban models that can effectively address and visualize these growing threats to avoid substantial loss of life and property. UVDAT: Urban Visualization and Data Analysis Toolkit bolsters data-driven research and aids the development and well-being of urban areas. The toolkit provides access to a comprehensive data library incorporating data science and AI methods for analysis of natural hazards on infrastructure. It also includes neighborhood-scale, high-resolution visualizations, for detailed data and uncertainty representations. During Phase I, we thoroughly researched and assessed methods to construct an efficient urban data platform capable of ingesting, processing, and transforming observational and simulation data. Key AI and data science techniques were identified to effectively merge and downscale climate model outputs for seamless integration with the data system. We explored diverse multi-scale visualization approaches and developed semi-automated analysis methods. These methods enable users to easily identify, compare, and interpret model output results. Our primary focus was the resilience of urban infrastructure, particularly targeting urban rail and subway systems at risk of flooding. In Phase II, our objective is to transform UVDAT into a comprehensive, turnkey solution, accompanied by web-based workflow interfaces tailored for urban decision-making processes. We will enhance our data library to handle larger and more diverse network and geospatial data, crucial for supporting these workflows. To facilitate real-time inference and model executions more effectively, enhancements to our integration APIs are planned. In terms of visualization, we aim to standardize data representation across different formats and introduce innovative techniques for easier comprehension of model outcomes. Furthermore, we intend to refine our downscaling products and flood modeling capabilities, and to broaden our application scope to include road transportation use cases. The proposed system is designed to assist urban planners, policymakers, and logistics engineers in managing increasingly large, complex, and diverse datasets. By integrating climate data with other critical data types such as infrastructure, sensor, and demographic information using our advanced analytic and visualization tools, planners will be empowered to make data-driven decisions.