Projects
Selected research and software projects.
A selection of projects in natural language processing, computational economics, simulation, and neuroeconomics. Each repository contains the code, documentation, and materials needed to understand the work.
View all projects on GitHubNatural language processing
Financial Text Modeling in Julia
An educational model suite that makes the core algorithms behind financial NLP explicit and inspectable.
About the project
This project reimplements a university exploration of financial-text analysis as a dependency-light Julia package. It covers the path from raw documents to sentiment scores, classifiers, topics, and word embeddings, while using a synthetic corpus so the examples can run without private data or API credentials.
What it includes
- Sentiment analysis and multinomial Naive Bayes classification
- TF–IDF, latent semantic analysis, NMF, and latent Dirichlet allocation
- Native Word2Vec, embedding evaluation, and SEC filing preprocessing
- Reproducible examples plus unit and integration tests
Computational social science
Crash Course Content Analysis
A reproducible analysis of how language and subject matter vary across educational-video transcripts.
About the project
The project turns a collection of Crash Course transcripts into a structured corpus and applies a consistent NLP workflow across series. It is designed to make comparisons interpretable rather than treating the transcripts as an opaque prediction dataset.
What it includes
- Corpus cleaning and descriptive text statistics
- Part-of-speech patterns and vocabulary comparisons
- Keyphrase extraction across subjects and series
- Repeatable analysis from source text to reported results
Operations and simulation
Online Food Delivery Simulation
A discrete-event model of order flow, service capacity, priority queues, and customer timeouts.
About the project
This project studies an online delivery system as a queueing problem. Orders move through a capacity-constrained service process, while priority rules and abandonment behavior shape waiting times, utilization, and the number of successfully completed orders.
What it includes
- Configurable arrival and service processes
- Priority-based queueing and customer abandonment
- Scenario comparisons for capacity and operating policies
- Synthetic examples that run without the original project data
Neuroeconomics
Narratives and Herding
A research project connecting narrative processing with herding behavior in economic decisions.
About the project
The project develops the idea that shared narratives can coordinate beliefs and decisions, helping explain why groups sometimes move together even when private information differs. It brings evidence from neuroscience and behavioral economics into a framework that can generate testable hypotheses.
What it includes
- Literature synthesis across narrative economics and neuroscience
- A proposed mechanism linking stories, attention, and coordinated choice
- Testable implications for herding behavior
- Supporting computational analysis in a documented notebook