What it does This repo is one long Jupyter notebook that predicts whether someone earns more than $50K/year using the UCI Census Income dataset. It runs through the standard sklearn hits: exploration, imputation, encoding, feature ranking, then trains models with both sklearn and TensorFlow. A companion mindmap (separate repo) maps out the broader data science workflow.

The interesting bit The Docker setup is almost aggressively simple—one docker run command and you’re at localhost:8888 with TensorFlow and Jupyter ready. For a field where environment setup eats half a day, that’s not nothing.

Key highlights

• Covers the full pipeline: univariate/bivariate exploration, imputation, selection, encoding, PCA, and model comparison
• Includes ROC curves and metric calculations (accuracy, precision, recall, f1) for algorithm comparison
• Designed to run on the official jupyter/tensorflow-notebook Docker image
• Companion mindmap/cheatsheet at dformoso/machine-learning-mindmap
• ~700 stars suggests it has served as a reference for others learning the sklearn workflow

Caveats

• The README is a walkthrough, not a library—expect copy-paste learning, not import-and-go code
• No requirements.txt or pip install instructions; Docker is the only documented path
• TensorFlow usage is mentioned but not detailed in the README; unclear if it’s a full alternative pipeline or a brief add-on

Verdict Good for someone who wants to see a complete, documented sklearn workflow in one place and prefers learning by running cells. Skip if you need modular, reusable code or are already comfortable building pipelines from scratch.