Research Projects

MacArthur Foundation Grant 106476-0: Urban Data Fusion for Modeling Urban Sustainability

PI:  Paul Waddell

In this project we are exploring ways to leverage emerging data science techniques to create a platform that enables the creation and maintenance of urban data to monitor and analyze urban development and transportation that is of improved quality and coverage, and which, to the maximum extent possible, can be shared as open data for planning and research. To meet these objectives we plan to broadly engage the data science community at Berkeley through interactions at BIDS and the Urban Analytics Lab, as well as leveraging tools such as the UrbanSim Cloud Platform. Our primary modeling platform is based on UrbanSim, an open source platform developed using Python and the scientific libraries available for it, including Numpy, Pandas, Statsmodels, as well as HDF5 and PostGIS for persistent data storage.

We are also evaluating means to radically speed up computation of these models and analyses, and are actively extending the models to other problem domains beyond our current choice models for household and firm location and real estate development, such as travel behavior, vehicle ownership, and adoption of energy and water-saving appliances. We are interested in potential use of GPU computing to radically speed up estimation, calibration, and simulation of these models for use in decision-support – and to be able to robustly represent the uncertainty in the inputs and predictions from the models. Finally, we want to make these models accessible to the public, in order to enable broader civic engagement in major decisions shaping the future of communities, such as land use regulations (zoning, comprehensive plans) and infrastructure (roads, transit, water, sewer, energy).

We will maintain open source repositories for the code developed within this project, on a GitHub repository, and will be coordinated with other open source projects that may offer valuable building blocks for the proposed platform.
Our objectives for year 1 of this project include assembly of the main data sources to be analyzed during the project, and an initial assessment of the issues in using and integrating them. Year 2 objectives will focus on developing the tools to improve the quality and completeness of the data, and the quality, robustness and computational performance of the models.

One of the initial results of this project is a compilation of rental listings data from across the United States, that will form a basis for ongoing analysis into housing markets dynamics and housing affordability.  This was published in the Journal of Planning Education and Research, and covered by the Washington Post.

California Air Resources Board, No. 13-310: Developing a New Methodology for Analyzing Displacement

PI: Karen Chapple, Co-PI: Paul Waddell, Daniel Chatman. In collaboration with UCLA PI: Anastasia Loukaitou-Sideris, and Co-PI Paul Ong

In 2008, California passed Senate Bill 375, requiring metropolitan planning organizations to develop Sustainable Communities Strategies as part of their regional transportation planning process. While the implementation of these strategies has the potential for environmental and economic benefits, there are also potential negative social equity impacts, as rising land costs in infill development areas may result in the displacement of low-income residents. This project examines the relationship between fixed-rail transit neighborhoods and displacement in Los Angeles and the San Francisco Bay Area, modeling patterns of neighborhood change in relation to transit-oriented development, or TOD. Overall, we find that TOD has a significant impact on the stability of the surrounding neighborhood, leading to increases in housing costs that change the composition of the area, including the loss of low-income households. We found mixed evidence as to whether gentrification and displacement in rail station areas would cause an increase in auto usage and vehicle miles traveled (VMT). The report also examines the effectiveness of anti- displacement strategies. The results can be adapted into existing regional models such as UrbanSim to analyze different investment scenarios.  The project included an off-model tool that will help practitioners identify the potential risk of displacement. 

NSF IIS-0964412: Integrating Behavioral, Geometrical and Graphical Modeling to Simulate and Visualize Urban Areas

PI: Paul Waddell, Co-PI: Michael Jordan, in collaboration with
Purdue University, PI: Daniel Aliaga, Co-PI Bedrich Benes

In this project, we developed a new simulation framework to interactively model and visualize socioeconomic and geometric characteristics of urban areas. The framework consists of a synergistic collaboration of three different areas: behavioral urban modeling, probabilistic graphical modeling, and visualization and computer graphics. In machine learning and statistics, the area of probabilistic graphical modeling offers a flexible framework to build, estimate and simulate from models of substantial complexity and scale, with partially observed data. By accounting for uncertainty and interdependencies, including aspects of dynamic equilibrium that arise in modeling the complex spatio-temporal dynamics of urban areas, we believe there is significant potential for breakthroughs in modeling large-scale urban systems.  Similarly, by integrating behavioral and geometrical dimensions of urban areas, we expect to exploit the power of behavioral simulations more effectively by filling in geometric details that behavioral models are not well suited to manage, and at the same time provide a powerful framework to generate 2D and 3D geometric representations of urban areas that are behaviorally and geometrically consistent. We are taking advantage of massive datasets available for urban areas, including parcel and building inventories, business establishment inventories, census data, household surveys, and GIS data on physical and political features, and fuse these data into a coherent and consistent database to support his modeling objectives. This data fusion will address imputation of missing data, accounting for complex spatial and relational connections among the data sources. 

This project, funded by the Metropolitan Transportation Commission, leveraged research from our NSF grant to implement a prototype of UrbanVision, a 3D visualization platform for exploring alternative future scenarios modeled using UrbanSim or created through a visioning process.  UAL was responsible for the development of the visualization platform used in outreach meetings for the PlanBayArea project.  As part of this project, UAL also developed an application of the UrbanSim model platform for the Bay Area, including data integration, model specification, estimation and calibration.  We worked closely with MTC and ABAG to analyze the alternatives developed as part of the Environmental Impact Report (EIR) process for the Sustainable Communities Strategies planning process mandated by state legislation SB375. Using UrbanSim in combination with the MTC activity-based travel model, MTC was the first Metropolitan Planning Organization in California to successfully complete a Sustainable Communities Strategy planning process using integrated land use and transportation models as mandated by SB375 and other state guidelines.