Abstract Submission: This study evaluates the impact of the 2017 Croton Water Supply change on water quality complaints in Manhattan, New York City. We used a Causal Forest (CF) model alongside a Difference-in-Differences (DID) approach, analyzing data from NYC311, Water Quality Reports, and the American Community Survey (2010-2024). The DID analysis identified a significant increase in complaints 2-3 years post-supply change, peaking in the third year (ATT = 18.50, p<br>< 0.05), indicating a delayed public response. The CF model, incorporating socioeconomic variables like income, population density, and building age, estimated heterogeneous treatment effects with an Average Treatment Effect (ATE) of 10.53, rising to 16.43 in the Post-Croton phase. Variable importance analysis revealed that the Gini coefficient, indicating income inequality, was the most influential factor in predicting treatment effects, suggesting higher responses in economically disparate areas. The CF model's partial dependence plots showed non-linear relationships between complaints and factors such as income, highlighting the role of local demographics in public response. The study underscores the need for proactive communication strategies in water management and provides a methodological framework for evaluating water supply changes in urban settings. By combining DID and CF analyses, the research offers a comprehensive view of the Croton Water Supply's impact, identifying key drivers of public response.
