A comprehensive system for downloading, transcribing, and analyzing the Tales From The East Stand (TFTES) Shamrock Rovers podcast archive.
This project processes the entire TFTES podcast archive (345+ episodes, 623+ hours) to create a searchable database with an interactive web interface. The system includes:
- Bulk Episode Downloader: Downloads all episodes from RSS feeds
- Audio Transcription: Uses OpenAI Whisper for high-quality speech-to-text
- Content Analysis: Rule-based classification for Shamrock Rovers topics
- Web Interface: Interactive dashboard with search and analytics
# Create virtual environment
python3 -m venv tftes_env
source tftes_env/bin/activate # Linux/Mac
# or
tftes_env\Scripts\activate # Windows
# Install dependencies
pip install requests feedparser openai-whisper torch pandas numpy# 1. Download episodes
python3 tftes_downloader.py
# 2. Transcribe audio
python3 tftes_transcriber.py
# 3. Analyze content
python3 tftes_analyzer.py
# 4. Serve web interface
python3 -m http.server 8000 # Then open index.htmltftesdotcom/
├── tftes_downloader.py # Downloads episodes from RSS feeds
├── tftes_transcriber.py # Transcribes audio using Whisper
├── tftes_analyzer.py # Analyzes content and creates database
├── index.html # Interactive web interface
├── tftes_episodes/ # Downloaded audio files
├── tftes_transcripts/ # Generated transcripts (JSON + text)
├── tftes_analysis/ # Analysis results and database files
└── tftes_env/ # Virtual environment
- RSS feed parsing with error handling
- Resume capability for interrupted downloads
- Metadata preservation (titles, descriptions, dates)
- Progress tracking and batch processing
- OpenAI Whisper with word-level timestamps
- GPU acceleration support
- Multiple output formats (JSON + text)
- High accuracy (>95%) for clear audio
- Shamrock Rovers-specific topic classification:
- Match Analysis
- Player Discussion
- League Standings
- European Competition
- Club Management
- Fan Culture
- History & Nostalgia
- Entity extraction (players, teams, matches)
- Sentiment analysis
- Searchable database generation
- Interactive episode browser
- Advanced search and filtering
- Analytics dashboard with charts
- AI chatbot for episode recommendations
- Mobile-responsive design
- Shamrock Rovers theming
- CPU: 4-core processor
- RAM: 16GB
- Storage: 100GB free space
- Network: Stable broadband
- CPU: 8+ core processor
- RAM: 32GB+
- GPU: NVIDIA RTX 3060+ (for faster transcription)
- Storage: SSD with 200GB+ free space
- Download Phase: 8-12 hours
- Transcription Phase: 60-80 hours CPU / 15-20 hours GPU
- Analysis Phase: 4-6 hours
- Database Creation: 2-3 hours
- Total: 3-5 days with optimal hardware
- Primary:
https://feeds.podcasts.apple.com/podcast/1209834670 - Alternative:
https://podcasts.apple.com/podcast/id1209834670?format=rss
- Total Episodes: 345+
- Total Duration: 623+ hours
- Average Episode Length: 108 minutes
- Total Word Count: 2.8+ million words
The web interface uses Shamrock Rovers brand colors:
- Primary green:
#008B44 - Light green:
#10B854 - Dark green:
#006633 - White/cream accents for contrast
- Sequential pipeline: download → transcribe → analyze → display
- Whisper model "large-v2" recommended for best quality
- GPU acceleration automatically utilized if available
- Rule-based analysis for consistency and cost-effectiveness
- Vanilla JavaScript with Chart.js for analytics
This project is for educational and personal use only. All podcast content remains the property of its original creators.
This is a personal project for Shamrock Rovers fans. Feel free to fork and adapt for your own podcast analysis needs.
For questions or issues, please refer to the implementation guide and code comments.