The Collapse of the American Dream: From "Rags to Riches" to the "Ragged and the Rich"

Capstone Project — Wharton Global Youth Data Science Academy
Authors: John Wu, Connie Lu, Smriti Vijay, Gwen Gao, Alvin Li

---

Project Overview

This project investigates how the promise of the American Dream has evolved over time and explores factors that influence social mobility. Using datasets from the US and international sources, we modeled the relationships between economic inequality, corruption perception, healthcare, education, and social mobility outcomes.

Key Research Questions:

• How accurately does the current American Dream provide upward mobility?
• How do economic and social policies affect social mobility across countries?
• What lessons can the U.S. learn from other nations to improve social mobility?

---

Repository Structure

data/
  WDS Project.xlsx      # Raw dataset used for analysis
  Clean_Data.csv        # Cleaned dataset

output/
  analysis.html         # Compiled HTML of analysis.Rmd
  report.html           # Compiled HTML of report.Rmd
  presentation.pptx     # Presentation slides

analysis.Rmd            # Full R Markdown with code, cleaning, and modeling
report.Rmd              # Clean narrative version for presentation
final.Rproj             # RStudio project file

README.md               # Project overview and instructions
REFERENCES.md           # List of all data sources and references

---

Methodology

1. Data Collection

   • Datasets from FRED, OECD, World Bank, Numbeo, Transparency International, Our World in Data, and more.

2. Data Cleaning & Preprocessing

   • Performed entirely in analysis.Rmd — no external “cleaned” CSV required.
   • Steps include subsetting, log-transformations, and variable selection.

3. Modeling

   • Techniques: Linear Models (LM), LASSO, Relaxed LASSO, Decision Trees, Random Forest, XGBoost
   • Predicting Social Mobility Index, Poverty Rate, and GDP per Capita

4. Model Selection

   • Selected based on testing error, interpretability, and statistical significance.

---

Key Findings

• Social Mobility Index: Influenced by corruption perception, wealth share of top 1%, and GINI coefficient.
• Poverty Rate: Affected by healthcare index, wealth distribution, and welfare spending.
• GDP per Capita: Strongly predicted by corruption perception.

See output/report.html for visualizations and detailed results.

---

Limitations

• Small sample size (24 countries) limits predictive power.
• Focused mainly on high-income countries; non-OECD countries are underrepresented.
• Some data points may be outdated.

---

Future Work

• Expand analysis to more countries and time periods.
• Examine real socio-economic effects of policy changes.
• Evaluate longitudinal trends in social mobility.

---

How to Run

1. Open final.Rproj in RStudio.
2. Open analysis.Rmd and knit to generate results.
3. Keep the raw dataset data/WDS Project.xlsx in the data/ folder.

---

References

• FRED, OECD, Our World in Data, Numbeo, Transparency International, World Bank, N26, World Population Review
• See References for more detailed sources.

---

Note: Rendered outputs in output/report.html may not knit perfectly on different machines due to package version differences. All code and preprocessing steps are reproducible via analysis.Rmd.