ericg1212/data-engineering-portfolio
Proprietary AI earns +111.5% higher Sharpe ratios than partnership AI. 5 Airflow pipelines (stocks, SEC EDGAR, FRED, crypto, weather) → AWS S3/Athena → Power BI. 163 pytest tests, Terraform IaC.
What's novel
Proprietary AI earns +111.5% higher Sharpe ratios than partnership AI. 5 Airflow pipelines (stocks, SEC EDGAR, FRED, crypto, weather) → AWS S3/Athena → Power BI. 163 pytest tests, Terraform IaC.
Code Analysis
14 files read · 4 roundsA production Airflow-based data pipeline that ingests stock prices (Alpha Vantage/yfinance), SEC 10-K capex data (EDGAR), and FRED macro indicators into S3/Athena, then computes Sharpe ratios, build-vs-rent AI premium, capex efficiency, and macro regime classification across 10 tech stocks to quanti
Strengths
Genuinely non-trivial financial analysis with correct Sharpe/drawdown/beta implementations, a well-structured 4-pipeline ETL architecture with proper S3 partitioning and Athena registration, data quality validation at each stage, and a sophisticated macro regime classification that tests the thesis across rate/inflation/unemployment conditions. Tests are meaningful with proper mocking of external dependencies and edge case coverage.
Weaknesses
The core thesis (builder vs renter categorization) is hardcoded in config.py rather than derived from data, the annualized return uses arithmetic mean compounding rather than geometric (minor methodological imprecision), the macro regime analysis has an O(n²) pattern in build_monthly_sharpe, and the Spearman correlation is computed on only 4 data points (the 4 AI_CAPEX companies) making the p-value claim statistically fragile.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
61
Contributors
1
Files
78
Active weeks
5
Repository
Language
Python
Stars
1
Forks
0
License
MIT