1,717 stars for someone else's homework answers
A repo of Jupyter notebooks solving 20 data-science take-home challenges from a paid book, minus the actual data.

What it does
This repository contains notebook-by-notebook solutions to the exercises in A Collection of Data Science Take-Home Challenges — a commercial book sold at datamasked.com. The author, Jifu Zhao, walks through 20 typical interview-style problems: conversion rates, A/B tests, fraud detection, funnel analysis, pricing tests, and so on. Each challenge gets its own .ipynb file with commentary and code.
The interesting bit
The README opens with a blunt disclaimer: “Please don’t contact me for the dataset.” The author won’t share the book or data, which makes the repo a curious artifact — popular enough to star, yet deliberately incomplete unless you’ve paid for the source material. It’s a study in how much value people find in seeing someone else’s worked examples even when they can’t run them.
Key highlights
- 20 notebooks covering bread-and-butter data-science interview topics (A/B testing, retention, clustering, credit risk, etc.)
- Clean, numbered structure:
01. Conversion Rate.ipynbthrough20. Ads Analysis.ipynb - Explicitly framed as self-learning material, not an official companion
- Links to a parallel community repo by stasi009 for additional reference
- Copyright 2018; no recent updates apparent from the README
Caveats
- No datasets included; you must purchase the book separately to reproduce anything
- The README is the entire documentation — no install instructions, requirements.txt, or environment notes
Verdict
Worth a skim if you’re preparing for data-science interviews and want to compare your approach to someone else’s. Skip it if you need runnable code or are looking for original challenges rather than solutions to a paid book.