Quantifying Intended and Unintended Impacts of Transportation Project Investments
In 2024, of the 830 projects approved under the $1.25 billion Dallas Bond program in the largest city in Dallas County, 57% or approximately $5.25 million were identified as “street and transportation.” Additionally, every two to three years, Dallas County Public Works partners with local municipalities to fund projects through the major capital improvement program (MCIP). Monies allocated through the MCIP fund transportation infrastructure projects that improve capacity and safety on the roads of unincorporated Dallas County (i.e., rural) and small urban municipalities. Over the next five years, projects including sidewalk improvements, street repairs, drainage improvements, and street rehabilitation will be funded across Dallas municipalities. The need to understand and manage local and regional impacts related to transportation investments has thus become a central focus. Specifically, concerns exist about impacts and risks beyond direct user costs and benefits. These include impacts such as visual and physical effects, land-use changes, local and regional shifts, losses to local businesses, impacts to economic productivity, as well as underinvestment. There is thus a desire to better understand and manage “intentional/beneficial” and “unintentional/adverse” project impacts and risks, and factors contributing to specific risks.
In collaboration with Dallas County municipalities, this research will develop a multi-criteria decision support tool and risk management framework for identifying and measuring impacts across project types and geographic scales. First, a national review of infrastructure project typologies will be developed to provide a baseline understanding of impacts. Guided by existing risk management frameworks, project impacts will be categorized into dimensions (e.g., physical, cyber, human, and economic) across micro, macro, and meso-scale projects. Mixed (i.e., quantitative and qualitative) data sources will be used to model project impacts. In an innovative approach, unsupervised natural language processing (NLP) will be used to thematically analyze and code public comments, gathered through the project development process. A database of impacts will be assembled by project typology and used to develop a risk management framework.
The resulting risk management framework will be applied in a case study analysis of selected infrastructure projects nationally. Statistical analysis will be used to identify trends associated with varying project typologies and impact categories based on historical data. This application will test the ability of the newly developed framework to identify unintended impacts and quantify local and regional risks within new contexts. Scenario analysis will also be used to link current-day investments and future risk profiles. Finally, recommendations will be made for policies that can inform future infrastructure investments and prioritization decisions.
The proposed project will deepen consideration of project development impacts by drawing on theories of risk management and measurement. Upon successful completion. This research will likely also lead to innovation in Dallas County infrastructure prioritization and investment strategies which are increasingly focused on being more risk-based. Anticipated benefits of the work include reduced costs from adverse and unintended project consequences. Finally, this approach will provide a bottom-up, stakeholder conceptualization of project impacts and risks, while advancing methods for processing public input data.
Project R2 will be completed in collaboration with cities across Dallas County, multiple departments and programs including the Department of Public Works, the Office of Information Technology, and the Bond Program. Local staff will provide detailed design information and monetary valuation of costs and benefits to support project typology development. Finally, local partners will contribute public comment data for NLP modeling and evaluation.
TBD