|
|
--- |
|
|
title: QualiInsightLab |
|
|
emoji: 🧪 |
|
|
colorFrom: green |
|
|
colorTo: blue |
|
|
sdk: streamlit |
|
|
sdk_version: 1.46.1 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
license: other |
|
|
short_description: AI-powered qualitative research transcription and analysis |
|
|
--- |
|
|
|
|
|
# 🧪 QualiLab |
|
|
## Advanced Qualitative Research Laboratory |
|
|
|
|
|
**QualiLab** is a comprehensive AI-powered platform for qualitative researchers, designed to streamline the entire research workflow from audio transcription to intelligent report generation. |
|
|
|
|
|
### 🎯 **What it does** |
|
|
|
|
|
QualiLab automates the most time-consuming aspects of qualitative research: |
|
|
|
|
|
1. **🎙️ Smart Audio Processing** |
|
|
- Automatic transcription using OpenAI Whisper |
|
|
- Supports multiple formats: MP3, WAV, MP4, M4A, AAC |
|
|
- Intelligent file splitting for large interviews (>25MB) |
|
|
- Handles both Focus Group Interviews (FGI) and In-Depth Interviews (IDI) |
|
|
|
|
|
2. **🤖 Intelligent Report Generation** |
|
|
- AI analyzes all transcriptions collectively |
|
|
- Generates comprehensive research reports |
|
|
- Self-prompting system ensures detailed, thorough analysis |
|
|
- Automatically detects interview type and adapts report structure |
|
|
|
|
|
3. **📊 Professional Output** |
|
|
- Executive summaries with key insights |
|
|
- Structured analysis with participant quotes |
|
|
- Business recommendations and next steps |
|
|
- Multiple export formats (TXT, with more coming soon) |
|
|
|
|
|
### 🚀 **Key Features** |
|
|
|
|
|
- **Multi-file Processing**: Upload and process multiple interviews simultaneously |
|
|
- **Large File Support**: Automatically splits files >25MB for processing |
|
|
- **Research Brief Integration**: Include study objectives for targeted analysis |
|
|
- **Real-time Progress Tracking**: Monitor transcription and report generation |
|
|
- **Cost Estimation**: See estimated API costs before processing |
|
|
- **Session Management**: Resume interrupted sessions |
|
|
- **Professional Reports**: Publication-ready research reports |
|
|
|
|
|
### 🛠️ **How to Use** |
|
|
|
|
|
1. **Setup**: Enter your OpenAI API key (requires access to Whisper and GPT-4o-mini) |
|
|
2. **Upload**: Add your interview audio/video files |
|
|
3. **Brief** (Optional): Describe your research objectives and target audience |
|
|
4. **Process**: Click "Start" and let QualiLab handle the rest |
|
|
5. **Analyze**: Review generated transcriptions and comprehensive report |
|
|
6. **Export**: Download your professional research report |
|
|
|
|
|
### 📋 **Requirements** |
|
|
|
|
|
- **OpenAI API Key** with access to: |
|
|
- Whisper API (for transcription) |
|
|
- GPT-4o-mini (for report generation) |
|
|
- **Supported File Formats**: MP3, WAV, MP4, M4A, AAC |
|
|
- **File Size**: Up to 200MB per file (automatically split if larger) |
|
|
|
|
|
### 💡 **Perfect for** |
|
|
|
|
|
- **Market Researchers** conducting focus groups and interviews |
|
|
- **UX Researchers** analyzing user feedback sessions |
|
|
- **Academic Researchers** processing qualitative data |
|
|
- **Product Managers** analyzing customer interviews |
|
|
- **Consultants** creating client reports from interviews |
|
|
|
|
|
### 🔬 **Technology Stack** |
|
|
|
|
|
- **Frontend**: Streamlit |
|
|
- **Transcription**: OpenAI Whisper API |
|
|
- **Analysis**: GPT-4o-mini with custom prompting |
|
|
- **Audio Processing**: Pydub, Librosa |
|
|
- **Deployment**: HuggingFace Spaces |
|
|
|
|
|
### 📊 **Sample Use Cases** |
|
|
|
|
|
**Focus Group Interview (FGI)** |
|
|
- Analyze group dynamics and consensus opinions |
|
|
- Identify different participant segments |
|
|
- Extract key themes and contradictions |
|
|
- Generate actionable business insights |
|
|
|
|
|
**In-Depth Interview (IDI)** |
|
|
- Deep dive into individual perspectives |
|
|
- Uncover personal motivations and barriers |
|
|
- Detailed journey mapping |
|
|
- Personalized recommendations |
|
|
|
|
|
### ⚡ **Performance** |
|
|
|
|
|
- **Transcription Speed**: ~10x faster than real-time |
|
|
- **Report Generation**: 5-10 minutes for comprehensive analysis |
|
|
- **Cost Efficient**: Typically $0.10-$0.50 per hour of audio |
|
|
- **Accuracy**: Professional-grade transcription and analysis |
|
|
|
|
|
### 🔒 **Privacy & Security** |
|
|
|
|
|
- Your audio files are processed securely through OpenAI APIs |
|
|
- No permanent storage of your data |
|
|
- API keys are handled securely within your session |
|
|
- Temporary files are automatically cleaned up |
|
|
|
|
|
### 🎓 **Tips for Best Results** |
|
|
|
|
|
1. **Good Audio Quality**: Clear audio produces better transcriptions |
|
|
2. **Detailed Brief**: More context leads to better analysis |
|
|
3. **File Organization**: Name files descriptively for easier tracking |
|
|
4. **Multiple Interviews**: Better insights when analyzing multiple perspectives together |
|
|
|
|
|
### 🆘 **Troubleshooting** |
|
|
|
|
|
**Large Files**: Files >200MB should be pre-compressed or split before upload |
|
|
**API Errors**: Ensure your OpenAI API key has sufficient credits and access |
|
|
**Processing Time**: Large files or multiple interviews may take 10-15 minutes |
|
|
|
|
|
### 🚧 **Roadmap** |
|
|
|
|
|
- [ ] Support for more export formats (DOCX, PDF) |
|
|
- [ ] Advanced analytics dashboards |
|
|
- [ ] Multi-language support beyond Polish/English |
|
|
- [ ] Integration with survey platforms |
|
|
- [ ] Sentiment analysis visualization |
|
|
- [ ] Participant persona generation |
|
|
|
|
|
### 📄 **License** |
|
|
|
|
|
**Copyright (c) 2025 Marek Staniszewski, www.Heuristica.pl. All rights reserved.** |
|
|
|
|
|
This software is proprietary. No part of this software may be reproduced, distributed, or transmitted without explicit written permission from the author. |
|
|
|
|
|
**For commercial licensing inquiries, contact:** [Contact via www.Heuristica.pl](https://www.heuristica.pl) |
|
|
|
|
|
**Academic and Research Use:** This tool is made available for educational and non-commercial research purposes. Users may utilize QualiLab for their research projects but may not redistribute, modify, or commercialize the software. |
|
|
|
|
|
--- |
|
|
|
|
|
**Built with ❤️ for the research community by [Heuristica.pl](https://www.heuristica.pl)** 🧪✨ |
