Assessing the impact of walkability indicators on health outcomes using machine learning algorithms: A case study of Michigan

Abstract

Urban planning and public health are increasingly interlinked in efforts to shape healthier communities. To build a healthier community, walkability has shown positive outcomes for population health. This study employs machine learning to analyze the impact of walkability indicators such as intersection density, proximity to transit stops, employment mix, and employment and household mix and social vulnerability factors on health outcomes in Michigan. Data from the Environmental Protection Agency (EPA) and Centers for Disease Control and Prevention (CDC) were used to evaluate health outcomes including obesity, blood pressure, cholesterol, and depression. The analysis also incorporated the Social Vulnerability Index (SVI) to examine the influence of socioeconomic and demographic factors. Different supervised machine-learning algorithms were applied to assess these relationships. Among the algorithms, the Random Forest algorithm showed the best performance. The results indicate that there is a variation in the impact of walkability indicators on health outcomes. Key findings reveal that among walkabiltity indicators, intersection density is the most significant predictor of all health outcomes, while the other indicators have less impact. In addition, it was found that variables such as Socioeconomic Status, Household Composition & Disability, Minority Status, Housing Type and Transportation have also impact of health outcomes. In conclusion, this research shows the relationship between walkability and human health by providing an evidence-based guidance for building healthier, more walkable communities.