Published Paper Details:

The health inequities across Indian districts have roots in Social Determinants of Health inclusive of maternal care, child nutrition, sanitation, education, and access to healthcare[4]. While datasets like NFHS-5 [1] are rich in information on these indicators, the high dimensionality coupled with inconsistent administrative formats causes serious challenge in systematic vulnerability assessment[14]. This paper presents HealthScape, a data-driven unified analytics framework that integrates PCAbased dimensionality reduction[14], [15], machine learning-based classification[6], [8], and geospatial visualization[7], [10] for generating district-level SDOH vulnerability maps. Standardized data cleaning and feature engineering, dimensionality reduction based on PCA, vulnerability scoring, and classification by XGBoost form the pipeline. District boundaries are spatially joined using GeoJSON files; hence, interactive choropleth maps can be generated to highlight high-risk clusters and regional disparities[15]. Experimental results confirm strong predictive performance and accurate spatial representation by identifying the top vulnerable districts and subsequently yielding interpretable insights for policymakers. Complex survey data will now be transformed into actionable intelligence, and the model proposed is scalable for population-health monitoring across India.