How mvmalyi/llm-semantic-maintenance-logs-analysis works
This is a reproducible research workflow for exploring semantic reliability insights in wind turbine maintenance logs. The documented user journey is to inspect data preparation, then run a processing notebook containing four LLM-based analyses. The repository also includes JSON, report, and timeline artifacts, but the README does not describe how they are produced or consumed. Relationships below are limited to those explicitly documented; notebook internals were not available for inspection.
Main entry points: 2_processing_log.ipynb Semantic analysis · 1_data_cleaning.ipynb Data preparation
Components 13
Analysis workflow
- Data preparation
1_data_cleaning.ipynb - Semantic analysis
2_processing_log.ipynb
Analysis capabilities
- Failure modes
2_processing_log.ipynb - Causal chains
2_processing_log.ipynb - Comparative analysis
2_processing_log.ipynb - Data quality audit
2_processing_log.ipynb
Repository artifacts
- Converter data
converter.json - Farm data
farms.json - Turbine data
turbine.json - Gemini quality report
data_quality_report_gemini.md - GPT quality report
data_quality_report_gpt.md - Interactive timeline
interactive_timeline.html
Other components
- Researcher
Connections 7
- Researcher → Data preparation: inspects
README.md - Data preparation → Semantic analysis: run sequentially
README.md - Researcher → Semantic analysis: runs
README.md - Semantic analysis → Failure modes: includes
README.md - Semantic analysis → Causal chains: includes
README.md - Semantic analysis → Comparative analysis: includes
README.md - Semantic analysis → Data quality audit: includes
README.md