Tools and Methods

The project utilizes a combination of technologies to clean, process, and analyze the data:

• Relational Big Data Management Systems (RBMDs): Custom RBMDs were created to store and manage large datasets efficiently, supporting both structured and unstructured data.
• SQL: SQL queries were used extensively to extract and manipulate data from the RBMDs, streamlining the data retrieval process.
• Python for Data Analysis and Visualization: Python libraries such as Pandas and NumPy were used for data preprocessing and manipulation, while Matplotlib and Seaborn were employed for creating visualizations, including heatmaps, dual-axis charts, and dashboards. These tools helped visually illustrate the relationships between key economic variables.
• BigQuery Integration: The data was processed and stored in BigQuery, facilitating seamless interaction with other tools and automating data flows into visualization platforms.
• Jupyter Notebooks: The entire analysis was performed in Jupyter Notebooks, making the workflow transparent and easy to reproduce.

Hypotheses and Findings

The analysis tests eight hypotheses, each supported by the data using methods like Vector Autoregression (VAR) modeling, correlation analysis, Ordinary Least Squares (OLS) regression, and Granger causality tests. The hypotheses are:

1. Hypothesis 1: An increase in nonfarm payroll employment (PAYEMS) is associated with a decrease in the unemployment rate (LNU02300000) over time.
2. Hypothesis 2: Periods of high employment (PAYEMS) correspond with higher Gross Domestic Product (GDP) growth rates.
3. Hypothesis 3: Increases in average hourly earnings (AHEMAN) lead to an increase in inflation (CPI).
4. Hypothesis 4: Wage growth in the manufacturing sector (AHEMAN) is positively correlated with overall wage growth across all sectors (AWHAE).
5. Hypothesis 5: The labor force participation rate (LNU02300000) decreases during economic recessions and increases during periods of economic recovery.
6. Hypothesis 6: Female labor force participation (LNS12000002) grows faster than male labor force participation (LNS12300001) during economic expansions.
7. Hypothesis 7: The growth rate of employment in the service sector (USSERV) is higher than in the goods-producing sector (USGOOD) over the last two decades.
8. Hypothesis 8: Manufacturing employment (MANEMP) has a negative correlation with automation indices over time.

All eight hypotheses are confirmed by the data, providing strong evidence of significant relationships between these economic indicators.

Visualization and Reporting

To make the analysis results more accessible and interpretable, the project includes:

• Interactive Dashboards: Built using Python, these dashboards allow users to explore the relationships between variables through visual tools like line graphs, bar charts, and heatmaps.
• Automated Data Pipelines: SQL scripts combined with Python automate data updates, ensuring that the visualizations and analysis stay current without manual intervention.

This project demonstrates advanced data analysis, programming, and visualization skills while offering deep insights into how U.S. economic trends evolve and influence each other. The findings not only confirm the hypotheses but also provide valuable context for understanding broader economic patterns, including how employment, wages, and GDP interact over time.