Study guide · 6 topics
Data Science Exam Study Guide
Six interactive topic guides — NumPy, Pandas, stats, ML fundamentals, model evaluation, and feature engineering. Identify your weak spots before your exam or assessment. Code examples, exam traps, and tips throughout.
Test your Data Science knowledge → Free · No sign-upNumPy Explained — Arrays, Broadcasting, Indexing & Common Operations
NumPy is the foundation of scientific Python. Here's what data science exams test — arrays, broadcasting, and vectorised operations.
Read guide →02 · 6 min readPandas Explained — DataFrames, Filtering, GroupBy & Merge
Pandas is the data manipulation library in Python. Here's what data science exams test — DataFrames, filtering, and aggregation.
Read guide →03 · 6 min readStatistics for Data Science — Mean, Distributions, Hypothesis Testing
Statistics is the language of data science. Here's what exams test — from descriptive stats to hypothesis testing.
Read guide →04 · 6 min readMachine Learning Overview — Supervised, Unsupervised & Key Concepts
Machine learning fundamentals are tested on every data science exam. Here's the core concepts and where beginners get confused.
Read guide →05 · 6 min readML Model Evaluation — Accuracy, Precision, Recall, F1 & ROC-AUC
Choosing the right evaluation metric is critical. Here's what exams test — precision vs recall tradeoff and when accuracy fails.
Read guide →06 · 5 min readFeature Engineering — Encoding, Scaling, Missing Values & Selection
Feature engineering often matters more than model choice. Here's what data science exams test.
Read guide →Where Data Science leads
Certifications that build on Data Science
Real proctored exams with published blueprints — each with a free readiness test here that scores you per exam domain rather than giving you one number.
Microsoft AI-103 (Azure AI Apps & Agents)
Three AI-103 domains — text analysis, information extraction and computer vision, up to 45% together — are the applied end of the work you just tested on.
NVIDIA NCP-AAI (Agentic AI)
Evaluation and Tuning is 13% of NCP-AAI, and evaluating a model against a rubric is a data-science problem before it is an agent problem.