​Cognitive Decline Predictors
Factors contributing to cognitive decline in older adults in the USA
Welcome to the Cognitive Decline Predictors Case Study. This project showcases my skills in coding, statistical analysis, and data visualization, and demonstrates how I translate complex data into actionable insights for stakeholders.
Project Overview
This project explored predictors of cognitive decline among older adults in the U.S. using data from the Behavioral Risk Factor Surveillance System (BRFSS). The analysis focused on how lifestyle behaviors — particularly physical inactivity and smoking — are linked to cognitive and overall health outcomes.
​
Goal: Identify actionable factors that could help prevent or slow the onset of cognitive decline and provide evidence to guide public health interventions.
Data & Methods
​Challenges & Solutions
-
Large dataset: Focused on key variables, excluded missing data.
​
-
Interrelated factors: Guided by prior research, highlighted smoking & inactivity.
​
-
Ensuring relevance: Chose analyses tied to policy & health promotion.
​
-
Communicating results: Used clear Tableau dashboards to reach non-technical audiences.
Key Findings
Figure 1. Geographic hotspots of smoking and physical inactivity among adults aged 50+ (BRFSS, 2015–2022). States with the highest rates include Mississippi, Kentucky, Arkansas, West Virginia, Oklahoma, Louisiana, and Tennessee.
Lifestyle Predictors
-
Lack of leisure-time physical activity and smoking strongly correlate with poor health outcomes and self-reported cognitive decline.
​
Trends Over Time
-
Smoking rates have declined steadily since 2011, showing policy success.
-
Physical inactivity rates have remained stagnant, highlighting a persistent challenge.
​
Geographic Hotspots
-
States with consistently high smoking and inactivity: Mississippi, Kentucky, Arkansas, West Virginia, Oklahoma, Louisiana, Tennessee
Recommendations
-
Promote physical activity (clubs, safe outdoor spaces).
-
Expand smoking cessation programs.
-
Integrate mental health + lifestyle programs.
-
Strengthen policy & track progress.
Skills Demonstrated
Python, Tableau, Excel · Data wrangling · Correlation & Regression · Geospatial mapping
Dashboard storytelling
Deliverables
Code Examples
​Code 1. Data wrangling: filter scope, keep “Overall” stratifications, and drop unused columns.
Code 2. Correlation heatmap: cognitive decline vs lifestyle and health indicators.
Code 3. Geographic hotspots — mapping state-level prevalence of smoking and inactivity using Folium choropleths.
Code 4. Multivariate regression — modeling poor health outcomes as a function of smoking and physical inactivity, with VIF check for collinearity.
