Portfolio Projects
A curated set of projects showing how I work with messy real-world data, geospatial analysis, OCR, computer vision, machine learning, and reproducible research workflows.
Flagship Case Studies
Start here if you want the clearest view of how I turn messy data into usable systems.
Research infrastructure that transforms historical university catalogs into structured datasets using scraping, OCR, extraction, and validation.
Proof: PDF -> OCR -> Dataset
Outcome: Created a reusable workflow for transforming archival catalogs into structured research data.
A geospatial decision-support workflow for identifying candidate utility-scale solar energy sites across Colorado.
Proof: Spatial Layers -> Suitability Score
Outcome: Generated a suitability map for utility-scale solar planning in Colorado.
Computer vision research workflow exploring how image enhancement can improve visual inputs for monocular SLAM systems.
Proof: Raw Frames -> Enhancement -> Feature Evaluation
Outcome: Built a repeatable workflow for comparing image enhancement methods.
Additional Projects
An AI interface concept for querying and exploring grid, energy, or infrastructure data through source-grounded responses.
Proof: Question -> Retrieval -> Grounded Answer
Outcome: Defined an AI system architecture for domain-specific data exploration.
A spatiotemporal research atlas mapping how economic geology education expanded across U.S. universities over time.
Proof: Archival Records -> Geocoding -> Atlas
Outcome: Created a geospatial dataset for historical education research.
A medical image classification project focused on model evaluation, error analysis, and responsible AI limitations.
Proof: Image -> Model -> Evaluation
Outcome: Built a chest X-ray classification workflow.
An imbalanced classification project focused on fraud risk scoring, precision-recall tradeoffs, and threshold selection.
Proof: Transactions -> Risk Score -> Threshold Decision
Outcome: Built an end-to-end fraud classification workflow.
Exploratory analytics project investigating ride demand patterns, time trends, and location-based transportation insights.
Proof: Ride Data -> Trends -> Dashboard
Outcome: Created visual summaries of ride demand patterns.