M.S. Applied Data Science - Capstone Chronicles 2025

10

this pattern. First, the composite lifestyle metric used here, while inclusive of multiple behaviors, may not fully capture the quality or intensity of those behaviors—meaning a “Good” score could still reflect suboptimal habits in clinically important areas. Second, lifestyle changes may require significant time before translating into measurable improvements in metabolic health, and cross-sectional data may fail to capture such lag effects. Finally, the persistence of high prevalence across all categories may indicate that clinical or genetic factors, as well as medication use, are exerting stronger influence than lifestyle alone. In the context of this study, these results support the idea that lifestyle measures may have limited power to discriminate between those with and without metabolic syndrome. Figure 3 Metabolic syndrome prevalence by lifestyle effort category (“Poor,” “Moderate,” and “Good”). Differences in syndrome prevalence were minimal across categories.

Income” (PIR ≤ 1.3), “Middle Income” (PIR between 1.3 and 3.5), and “High Income” (PIR > 3.5). These thresholds align with U.S. federal definitions used in nutrition and health policy research, where values near 1.3 typically correspond to eligibility for federal assistance programs, and values above 3.5 reflect higher relative affluence. As shown in Figure 4, the prevalence of metabolic syndrome was again approximately 60% in all three income groups, closely mirroring the lifestyle effort results. This suggests that, within this dataset, income level alone is not a strong differentiator of metabolic syndrome risk. Possible explanations include the mediating effects of lifestyle and clinical factors, the inability of cross-sectional data to capture long-term socioeconomic impacts, and environmental or cultural influences that affect the population broadly, regardless of income. Figure 4 Metabolic syndrome prevalence by income level (Low, Middle, High). All groups show similar proportions of affected individuals.

Following the analysis of lifestyle effort, socioeconomic status—measured by the income-to-poverty ratio (PIR)—was examined for its potential relationship to metabolic syndrome prevalence. For comparability, PIR values were binned into three categories: “Low

Given that neither lifestyle effort nor income level showed substantial variation in prevalence,

161