Data Analytics · Public Health
COVID-19 Data Analytics for Alabama Counties
Exploratory analysis of county-level COVID-19 data for Alabama, focusing on death rates, vaccination coverage, and provider distribution to understand which regions were most impacted.
Overview
This project explores COVID-19 outcomes across all Alabama counties using real-world case, death, testing, vaccination, and provider data. The goal was to clean and combine multiple datasets, engineer meaningful metrics (like death rate and vaccination rate), and visualize differences between counties.
The notebook walks through data cleaning, feature engineering, and exploratory visualizations, and can be extended into a foundation for simple dashboards or further modeling.
Data & Methods
Datasets
- County-level COVID-19 cases, deaths, and tests for Alabama.
- Vaccination data for population 16+ by county.
- Provider dataset listing vaccine locations by county.
Key Steps
- Loaded CSV files into pandas and removed aggregate or invalid rows.
- Converted numeric fields (cases, deaths, population, doses) to proper types.
- Engineered features such as:
death_rate = (deaths / cases) * 100vaccination_rate = fully_vaccinated / population_16_plus- Provider counts per county.
- Merged datasets on county name to align outcomes and access metrics.
Tech Stack
Python, pandas, NumPy, matplotlib, seaborn, Jupyter Notebook
Key Metrics & Findings
- Calculated county-level death rates (%) and ranked counties by severity.
- Computed vaccination coverage for population 16+ and identified counties with low coverage.
- Counted vaccine providers per county and compared them with vaccination and death rates.
- Highlighted counties with high death rate and low vaccination rate as potential areas of concern.
You can customize this section with specific numbers (e.g., “Top county death rate was X%”) once you finalize the analysis.
Key Charts
A few of the visualizations exported from the notebook:
Export your charts from Jupyter as PNGs and place them in assets/ to use these slots.
Challenges & Learnings
- Handling inconsistent or missing county names when merging datasets.
- Ensuring rate metrics were computed using the correct denominators (e.g., population 16+).
- Choosing chart types that clearly communicate relative differences between counties.
- Gained practice in working with real, messy public health data instead of toy datasets.