Data Science Coding Challenge - Solved!

In this video, I walk through one of the most common data science coding challenges you’ll encounter in interviews and on the job: analyzing an experiment from raw customer visit and order data. At first glance, it looks straightforward — you have a dataset with visit dates, an indicator for whether customers placed an order, and the order amount (which can also be zero). The task seems simple: evaluate the experiment and report results.

But here’s the catch — there are actually two subtle “gotchas” hiding in this problem. Missing them can completely change your conclusions about the experiment’s success or failure. These aren’t just small coding details; they represent real-world pitfalls in experiment analysis that can distort results and lead to wrong business decisions.

For data science career and interview coaching: https://www.whatstheimpact.com/services/

In this tutorial, I’ll:

Show you the Python code to analyze the experiment step by step

Explain the two key mistakes that can mislead your analysis

Demonstrate how correcting them dramatically changes the outcome

Highlight why these kinds of issues are so important in real A/B testing and experiment analysis for tech companies

This is more than just a coding exercise. It’s a lesson in how subtle data issues can have massive consequences, whether you’re evaluating marketing campaigns, product feature launches, or customer behavior. These are the kinds of challenges you’ll see in data science interviews, especially at top tech companies where experiments drive decision-making.

If you’re preparing for a data science coding interview or want to improve your skills in experimentation and causal inference, this video will give you both the code and the intuition you need to succeed.

📌 Topics covered:

Experiment analysis in Python

Data cleaning and validation for A/B tests

Common pitfalls and hidden gotchas in experimental data

Difference between “naive” analysis vs. robust analysis

Why interviewers use this question to separate strong candidates from the rest

Watch until the end to see the full solution and the lessons you can apply in both interviews and real-world projects.

#datascience #datascientist #datasciencetutorial #datascienceinterview #datacodingchallenge #techinterview #abtesting #experimentanalysis #python #codinginterview

Видео Data Science Coding Challenge - Solved! канала Jonathan.Interviews