This repository contains various coding experiments and projects created with the assistance of Claude AI. Each branch represents a different experiment or project, showcasing different technologies, frameworks, and use cases.
A Python package for generating NBA-style hexbin shot charts, similar to those popularized by Kirk Goldsberry. Features include:
- Hexbin visualization of shot data
- Support for NBA API data fetching or synthetic demo data
- Customizable gridsize, color modes, and minimum shots per hex
- Output to PNG files
Tech Stack: Python, Matplotlib, NumPy, Pandas, NBA API (optional)
A modern, responsive Progressive Web App for personal finance tracking. Built as a PWA with offline-first architecture.
- Manual transaction entry and receipt scanning with OCR
- Smart category detection and multi-currency support
- Analytics dashboard with charts and spending predictions
- Budget setting with progress tracking
- Export/import functionality
- Dark mode and PWA installation support
Tech Stack: React, TypeScript, Vite, Tailwind CSS, Tesseract.js, Recharts, Workbox
A fast, zero-configuration CLI tool for converting files between various formats. Supports images, documents, data files, and videos.
- Single file and batch directory conversions
- Image formats: PNG, JPG, WebP, BMP, GIF, TIFF, ICO, HEIC/HEIF
- Document formats: Markdown to HTML/PDF/DocX, HTML to PDF
- Data formats: CSV, JSON, YAML, XLSX, TSV conversions
- Video formats: MP4, AVI, MKV, MOV, WebM, FLV, WMV, M4V, TS
- Quality and resize options
Tech Stack: Python, Click, Pillow, Pandoc, OpenPyXL, PyYAML, ffmpeg
A comprehensive flight price monitoring system focused on routes from Basque Country airports (BIO, EAS, VIT) to European and transatlantic destinations.
- Multi-provider support: Kiwi Tequila API, Amadeus, Google Flights via SerpAPI
- Smart deal detection comparing current prices to historical averages
- Flexible date searches and cabin bag filtering
- Watch mode for specific routes with price drop alerts
- Web dashboard with price trend charts
- Telegram notifications and email reports
- SQLite database for historical price tracking
Tech Stack: Python, Flask, Chart.js, SQLite, Telegram Bot API, SMTP
A full-stack personal travel blog application with AI-powered insights and optional Office 365 integration.
- Travel documentation with photo galleries and interactive maps
- AI analysis for restaurants, recommendations, and highlights extraction
- Search and filtering by destination, favorites, timeline view
- OneNote and OneDrive integration for importing content
- JWT-based admin authentication
- Responsive design with Docker deployment options
Tech Stack: Node.js, Express, React, TypeScript, MongoDB, OpenAI API, Microsoft Graph API, Leaflet, Docker
A modern, responsive personal website for Stefano Masneri, Senior AI Engineer. Features include:
- Interactive neural network visualization with Three.js
- Sections for about, research, publications, skills, blog, and contact
- Responsive design with smooth scrolling navigation
- Auto-geocoding and globe visualization
- Test suite for functionality
Tech Stack: HTML, CSS, JavaScript, Three.js, Leaflet
A football match prediction engine for Serie A, Premier League, La Liga, and Bundesliga using statistical modeling.
- Poisson distribution model with Dixon-Coles correction
- Expected Goals (xG) data integration from Understat
- Live bookmaker odds integration via The Odds API
- Value bet detection using expected value calculations
- Kelly Criterion stake sizing
- Multi-league support with team form and head-to-head statistics
- Web interface with best bets panel
Tech Stack: Python, Flask, HTML, CSS, JavaScript, football-data.org API, The Odds API, Understat
Image classification and face recognition pipelines using machine learning.
- Image classification into categories: food, landscapes, people
- Multi-label face recognition (identifying multiple people in one image)
- Training scripts for custom classifiers
- Zero-shot classification using CLIP
- FastAPI web API for inference
- Face embedding enrollment and prediction
Tech Stack: Python, FastAPI, scikit-learn, CLIP, face_recognition, Uvicorn
Each experiment is contained in its own branch. To explore a specific project:
-
Clone the repository:
git clone https://github.com/Stocastico/Claude-code-experiments.git cd Claude-code-experiments -
Switch to the desired branch:
git checkout <branch-name>
-
Follow the setup instructions in the branch's README.md file.
This repository serves as a collection of experimental projects. Each branch is self-contained and may have its own contribution guidelines.
See LICENSE file for details. /Users/fasteno/Documents/CODE/Claude-code-experiments/README.md