🧹
Data Cleaning Bench
Spot duplicate rows, missing values, and inconsistent formats across five realistic messy business datasets before they reach a model.
Model Training, Evaluation & AI Ethics Simulator
Learn practical data science by doing it. Clean a messy business dataset, train and tune a model without overfitting it, read a confusion matrix like an analyst, and audit a deployed model's outcomes for bias before it ships.
Small business owners, new graduates entering data or analyst roles, and anyone who wants to understand what a model is actually doing before trusting it.
Spot duplicate rows, missing values, and inconsistent formats across five realistic messy business datasets before they reach a model.
Pick classification or regression, tune a complexity slider, and watch train/test accuracy respond — including the overfitting trap.
Given real TP/FP/TN/FN counts, compute precision, recall, and F1 — the metrics that matter beyond raw accuracy.
Diagnose disparate impact in a deployed model's outcomes across groups, then pick the fix that's actually proportionate to the evidence.
40 quiz questions across data fundamentals, ML basics, evaluation metrics, and AI ethics — each with a full explanation.
Earn Data Janitor, Model Trainer, Metrics Analyst, and Fairness Auditor badges, climb the leaderboard, and unlock a completion certificate at 100%.
Completely self-contained sandbox — no internet required when run locally.
Educational simulation only — all datasets, models, and audit scenarios are fictional and simulated locally. No real customer data is ever used.
Every session pairs with a quick 15–20 min real-world task that turns the simulator into a habit.
Look at any dashboard or report you use at work, and list one metric on it that could be misleading without more context (sample size, time period, or definition).
Data Science & ML Lab is part of our Business & Workforce Track.
No payment required to apply. We'll contact you to confirm access options.