Analytics Day 0 Cheat Sheet

Everything you need to recall in the first 60 seconds of an analytics interview - core frameworks, key formulas, SQL execution order, and the 10 golden rules. Use this as a Day 0 recall sheet when the interviewer expects structured thinking before deeper analysis.

• SQL Execution Order: FROM & JOIN, WHERE, GROUP BY, HAVING, SELECT, ORDER BY, LIMIT.
• Analytics Framework - Quick Reference: CRISP-DM, Hypothesis Testing, A/B Testing, Cohort Analysis, RFM Segmentation, Funnel Analysis.
• Key ML & Statistics Metrics: Accuracy, Precision, Recall, F1 Score, RMSE, R².
• Confusion Matrix 2×2: True Positive (TP), False Positive (FP), False Negative (FN), True Negative (TN).
• A/B Testing Framework: Define Hypothesis, Calculate Sample Size, Run Experiment, Collect Data, Statistical Significance, Business Decision.
• Bias-Variance Trade-off: underfitting is high bias; overfitting is high variance.
• 10 Golden Rules: clarify the business problem, state assumptions, think in metrics, caveat correlation versus causation, segment before averaging, and end with a recommendation.

How to Use This Cheat Sheet

This recall sheet moves from query logic to analytics frameworks, then to model metrics, experiment flow, diagnostics, formulas, and interview rules. The big picture is simple: first understand how the data is produced and filtered, then choose the framework, metric, diagnostic, and recommendation.

Analytics Framework - Quick Reference

Use the framework table to quickly decide what kind of analytics answer the case needs. CRISP-DM is the Cross Industry Standard Process for Data Mining and is used for an end-to-end analytics project.

Key ML & Statistics Metrics

These metrics help you choose the right evaluation lens for classification and regression problems. The main interview cue is the use case: balanced classes, high false positive cost, high false negative cost, imbalanced dataset, regression error, or variance explained.

Confusion Matrix 2×2

The confusion matrix gives the base vocabulary for classification metrics. It separates correctly predicted Positive and Negative cases from Type I Error - False alarm and Type II Error - Missed case.

A/B Testing Framework

A/B Testing is used for comparing variants, with Lift % and significance as the key output. The flow starts with the hypothesis and ends with a business decision.

Bias-Variance Trade-off Visual

The bias-variance trade-off links model complexity with training error, test error, and the problem type. The optimal model balances bias and variance with low training and test error.

Underfitting = Model too simple; misses patterns in training data (high bias).

Overfitting = Model memorises training data; fails on new data (high variance).

Solution: Regularisation (Ridge/Lasso), cross-validation, pruning, more data.

Key Formulas

These formulas are the Day 0 recall set for classification, regression, error measurement, variance explained, and retention strategy. Use them exactly when the metric matches the business or modelling question.

10 Golden Rules for Analytics Interviews

1. Always clarify the business problem before diving into data.
2. State assumptions explicitly - interviewers value structured thinking.
3. Think in metrics: what are we optimising? What's the North Star?
4. SQL is your first language - know JOINs, window functions, CTEs cold.
5. Correlation ≠ Causation - always caveat observational insights.
6. Start simple (mean, median, distribution) before advanced models.
7. A/B test everything - know sample size and duration calculation.
8. Segment before you average - Simpson's Paradox is everywhere.
9. Communicate findings visually - charts > tables > paragraphs.
10. End with a recommendation, not just an analysis.

The most frequent error is jumping into data, SQL, or advanced models before clarifying the business problem. Always clarify the business problem, state assumptions explicitly, think in metrics, and end with a recommendation, not just an analysis.