Research Projects & Portfolio

Here are some of my key research projects and applications, demonstrating expertise in applied economics, machine learning, and data science.

Sentiment Volume Forecasting

Machine Learning for Financial Markets

Research project investigating the usefulness of news sentiment in predicting intraday stock trading volume for U.S. airline stocks. This comprehensive study demonstrates:

  • Scale: Analysis of 1.3 million news articles from GDELT database
  • Real-time Processing: 15-minute increment predictions during trading hours
  • Advanced ML: TensorFlow, PyTorch, and ensemble methods
  • Performance: Results comparable with state-of-the-art literature
  • Period: January 2018 to May 2025 data coverage

Technologies: Python, TensorFlow, PyTorch, pandas, GDELT API, real-time data processing

Key Finding: While sentiment features provide marginal improvement in forecasting accuracy, the framework demonstrates the potential for real-time business monitoring applications.

Semantic Document Indexing Tool

Advanced semantic indexing application for PDF repositories using state-of-the-art natural language processing. Features include:

  • Semantic Search: Beyond keyword matching using transformer models
  • Large-scale Processing: Efficient handling of document collections
  • NLP Pipeline: Advanced text processing and embeddings
  • User Interface: Simple GUI for search and retrieval

Technologies: Python, sentence-transformers, Hugging Face, NLP, document processing

Application: Enables intelligent document search for research and business applications.

Labor Economics Research

Applied Econometric Analysis

Comprehensive econometric analysis examining labor market dynamics and policy impacts using modern microeconometric methods.

Focus Areas:

  • Employment and wage determination
  • Policy evaluation using difference-in-differences
  • Causal inference in labor markets
  • Statistical modeling and hypothesis testing

Technologies: R, econometric modeling, statistical analysis, data visualization

Development Economics

Economic Development Analysis

Research on economic development topics using quantitative methods and empirical analysis.

Approach:

  • Jupyter notebook-based analysis
  • Modern econometric techniques
  • Policy-relevant research questions
  • Data-driven insights

Technologies: Jupyter Notebooks, Python, statistical analysis, economic modeling

Forecasting & Predictive Analytics

Forecasting Projects

Predictive Analytics Portfolio

Multiple projects demonstrating expertise in forecasting and predictive modeling:

  • Traditional Methods: ARIMA, exponential smoothing, time series analysis
  • Machine Learning: Neural networks, gradient boosting (LightGBM)
  • Ensemble Methods: Model combination and performance optimization
  • Real-world Applications: Business forecasting and decision support

Technologies: R, Python, neural networks, LightGBM, time series analysis

Research Replication

Academic Research Standards

Replication materials for academic research demonstrating:

  • Methodological rigor and reproducibility
  • LaTeX document preparation
  • Academic writing and presentation
  • Research transparency and open science practices

Technologies: LaTeX, academic research methods, statistical replication

Technical Tools & Utilities

Stay Awake

Python Utility

Simple Python script to keep computers awake during long-running processes. Demonstrates practical problem-solving and tool development skills.

Technologies: Python, system utilities, automation

Research Philosophy

My projects reflect a commitment to:

  • Methodological Rigor: Applying appropriate statistical and econometric methods
  • Reproducibility: Open-source code and transparent research practices
  • Real-world Relevance: Addressing practical problems with academic rigor
  • Technical Excellence: Using state-of-the-art tools and methods
  • Interdisciplinary Approach: Combining economics, computer science, and statistics

Future Directions

I continue to work on projects at the intersection of economics and data science, with particular interest in:

  • Real-time economic monitoring using big data
  • Machine learning applications in policy evaluation
  • Advanced econometric methods for causal inference
  • Open-source tools for economic research

All projects are available on GitHub with open-source licenses encouraging reuse and collaboration.