Digital Food Trend Analysis
Can data tell the difference between a viral moment and a lasting cultural shift? Using 10 years of YouTube, Reddit, and Google Trends data, I mapped how food trends are born, travel, and die, and what signals separate the fleeting from the permanent.
Tags: Python · NLP · LDA · Sentiment Analysis
Financial Market & Performance Analysis
74 financial variables. One stock. I engineered a feature selection pipeline that cut through the noise to build a model explaining 98% of Citigroup's daily price movements.
Tags: Python · SAS · Regression · Feature Engineering
Computer Vision & Age Detection
A 19% accuracy improvement not from a better model, but from rethinking how the problem was framed. Restructured 11 age classes into 5 demographic clusters across 20,000 facial images.
Tags: VGG16 · Transfer Learning · PCA · PyTorch
NYC Airbnb Market Analysis
Is a $15,000 renovation worth it? Using Tableau and geospatial pricing data across NYC boroughs, I answered that question with a number: 49 rental nights to break even.
Tags: Tableau · Mapbox · Geospatial · ROI Analysis
Airline Junk Fee Anatomy
Which airlines make the most money from fees nobody asked for and does consumer backlash actually change anything? An end-to-end data engineering pipeline using LLMs to extract structure from thousands of raw DOT complaint narratives, loaded into Snowflake and visualized in Tableau.
Tags: Python · Snowflake · dbt · LLM Extraction · Tableau
-Work In Progress-