How Data Scientists Build Interpretable ML Models with SHAP

What this episode covers

Lucas and Luna explore the practical use of SHAP (SHapley Additive exPlanations) for interpreting complex machine learning models. They walk through a real-world example: a credit risk model from a mid-sized European fintech that needed regulatory compliance under GDPR. Lucas explains how SHAP values decompose a prediction into feature contributions, and why game theory provides a principled foundation. Luna questions whether SHAP is always better than simpler alternatives like LIME, and they compare trade-offs in speed, consistency, and trust. The episode includes a concrete walkthrough of a single prediction breakdown, showing how a 32-year-old applicant with a thin credit file got denied because of a specific feature interaction. They also touch on open-source tools like the SHAP Python library and how one data team at Klarna uses SHAP summaries to communicate with non-technical stakeholders. No clickbait, just a clear look at one of the most widely adopted interpretability methods in the field today. #SHAP #InterpretableML #ExplainableAI #XAI #SHAPValues #GameTheory #FeatureImportance #ModelInterpretability #CreditRiskModeling #GDPR #LIME #Klarna #DataScience #Technology #MachineLearning #FexingoBusiness #BusinessPodcast #DataDriven Keep every episode free: buymeacoffee.com/fexingo

PodParley-generated summary based on available episode metadata and transcript content.