📈 EDA PROJECT 5:

Unveiling Mortality Trends: A Data-Driven Exploration of Leading Causes of Death in the United States (1999–2017)

📌 Project Overview

This project involves my comprehensive exploratory data analysis (EDA) on the NCHS Leading Causes of Death dataset for the United States. I obtained the dataset from the National Center for Health Statistics (NCHS) via DATA.GOV, which provides mortality data by cause of death, state, and year for the period 1999-2017.

My primary objective in this analysis is to gain insights into the patterns and trends of leading causes of death in the United States, including:

• Identifying the most prevalent causes of death.
• Examining how death counts and age-adjusted death rates vary across different states and over time.
• Understanding the relationship between raw death counts and age-adjusted death rates.
• Exploring changes in the ranking of leading causes over the years. Through this EDA, I aim to uncover key findings that can contribute to a better understanding of mortality in the US and potentially inform public health strategies. Furthermore, this project is part of my public data-science portfolio and showcases reproducible EDA practices, clear visual storytelling, and careful interpretation of population-level health data.

---

🧰 Tools & Libraries

• Python (Jupyter Notebook)
• pandas, numpy for data wrangling
• matplotlib, seaborn, plotly for visualization
• Git & GitHub for version control and project hosting

---

❓ Some of the Key Questions Explored

This analysis was guided by 12 questions designed to uncover timely and actionable patterns in U.S. mortality. Here are seven of those:

1. Which cause of death has the highest total number of deaths across all years and states?**
2. Which cause of death has the lowest total number of deaths across all years and states?
3. What is the distribution of deaths across different states for a given year?
4. What is the average age-adjusted death rate for each cause of death?
5. What is the trend of total deaths in the United States over the years?
6. How does the number of deaths for a specific cause vary by state?
7. How has the age-adjusted death rate for a specific cause changed over time?

---

📊 Key Findings and Implications

Based on our comprehensive exploratory data analysis of the NCHS Leading Causes of Death dataset, here are some key findings and their potential implications:

Key Findings:

• Dominance of Heart Disease and Cancer: Consistent with national trends, Heart disease and Cancer are the leading causes of death among the specific causes listed. In 2017, Heart disease accounted for approximately 61.0% and Cancer accounted for approximately 56.4% of total deaths from these specific causes (as shown in the answer to Question 15). This highlights the ongoing public health challenge posed by these diseases.
• Increase in Total Deaths Over Time: The overall trend of total deaths in the United States shows a general increase from 1999 to 2017 (as shown in the answer to Question 5). This could be attributed to various factors, including population growth, an aging population, or changes in mortality rates for specific causes.
• Variability Across States: Death counts and age-adjusted death rates vary significantly across states. For example, in 2017, California, Florida, and Texas had the highest total death counts, while Alaska had the lowest (as shown in the answer to Question 3). This underscores the importance of state-level analysis to understand regional differences in mortality patterns.
• States with Consistently Higher/Lower Rates: Our analysis identified states that tend to have consistently higher or lower age-adjusted death rates for specific causes over the years (as shown in the answer to Question 11). For instance, Wyoming had the highest average age-adjusted death rate for both CLRD and Suicide (as shown in the answer to Question 8), while states like New York had the lowest average rates for Alzheimer's disease and Stroke (as shown in the answer to Question 9). This information can be valuable for targeted public health interventions and resource allocation to address disparities.
• Difference Between Raw Counts and Age-Adjusted Rates: The notable difference between raw death counts and age-adjusted death rates, particularly in populous states and for high-incidence causes (as shown in the answer to Question 12), emphasizes the importance of using age-adjusted rates for comparing mortality risk across populations with different age structures.
• Relative Stability in Top Cause Ranking: While the number of deaths has fluctuated, the relative ranking of the top leading causes of death has remained largely consistent over the years (as shown in the answer to Question 14).

Implications:

• The persistent high prevalence of Heart disease and Cancer necessitates continued focus on prevention, early detection, and treatment strategies for these diseases.
• The increasing trend in total deaths highlights the need for ongoing monitoring of mortality rates and further investigation into the contributing factors.
• The significant variations in death rates across states suggest that public health initiatives may need to be tailored to address state-specific challenges and risk factors.
• Identifying states with consistently higher or lower rates for certain causes can help prioritize public health efforts and facilitate the sharing of best practices.
• Using age-adjusted death rates in public health reporting and analysis is crucial for accurate comparisons and informed decision-making, avoiding misleading conclusions based solely on raw death counts.

---

📸 Visuals

Total Deaths vs Sum of Specific Causes	Avg Age-adjusted Death Rate (Top10)

Top 10 States by Heart Disease Deaths	Trend: Age-adjusted Death Rate (Heart Disease)

Trend: Total Deaths for Top 5 Causes	Trend: Total Deaths in the U.S. Over the Years

---

📁 Files in This Repository

• EDA_Project_5_US_Mortality_Trends.ipynb — Jupyter notebook with full data pipeline, EDA code, tables, and figures
• images/ — plot PNGs used in the README and analysis
• data/ — The raw NCHS data are included here. The link is also available below (See Dataset Access & Sources).
• README.md — this file

---

📂 Dataset Access & Sources

Primary data sources and references used in this analysis (add or replace with the exact source links you used in the notebook):

• NCHS — Leading Causes of Death datasets (publicly available via DATA.GOV)

---

🚀 Future Works

1. Statistical testing: test year-to-year changes and trend breakpoints using time series tests (e.g., Mann-Kendall, segmented regression).
2. Modeling: build short-term forecasting models for top causes (e.g., ARIMA or Prophet), and consider cause-specific modeling for policy simulation.
3. Geospatial analysis: map state-level rates to identify clusters and hotspots.

---

✍️ Author

Adebayo Fashina 📍 Toronto, Canada | GitHub | LinkedIn

---