Die Küchenschlacht Analysis

Analyzing gender patterns in German cooking show using automated NLP pipeline

Automated data extraction pipeline analyzing gender patterns in 200+ episodes of the German cooking show “Die Küchenschlacht” (The Kitchen Battle).

Summary

  • Goal: Test whether juror/candidate gender affects competition outcomes
  • Method: MLX Whisper transcription → GPT-4 extraction → Statistical analysis
  • Sample: 200+ episodes, 1000+ candidates (2023-2025)

Key Findings

  • No evidence of gender bias (female jurors don’t favor female candidates)
  • Tasting order matters: Earlier positions have advantage
  • Effect weaker for female jurors

Tech Stack

Python • MLX Whisper • OpenAI GPT-4 • Pandas • Statsmodels

View full analysis and code on GitHub