Free tool · No sign-up needed
5 free, original PBQs for DataAI (DY0-001), with drag-and-drop. Do each task, check it, and read why the answer is right.
Last updated
CompTIA DataAI
You do each task instead of picking an answer. Every question is explained after you check it, and you get a score at the end.
Take it further in OpenpassAI
OpenpassAI’s practice sessions add performance-based questions, with command-line and network topology tasks, to full timed exams, AI explanations and a readiness score.
The questions
Each one with its scenario and task. Try it above first, then open the explanation.
Question 1 · Sorting · Mathematics and statistics
Before choosing statistical tests for a health survey, you need the level of measurement of each variable, because it decides which summaries and tests are valid.
Put each variable under its level of measurement.
Nominal values are names with no order. Ordinal values have an order, but the gaps between them aren’t equal or known. Interval values have equal gaps but no true zero: 20 °C is not twice as hot as 10 °C. Ratio values have equal gaps and a true zero, so ratios mean something: 180 cm is twice 90 cm. A mean is valid for interval and ratio data, and not for nominal data.
Question 2 · Matching · Modeling, analysis, and outcomes
A fraud model was tested on 1,000 transactions. It flagged 100 as fraud, and 80 of those really were. It missed 10 real frauds, and correctly passed 890 legitimate transactions. So there were 80 true positives, 20 false positives, 10 false negatives and 890 true negatives.
Match each metric to its value.
Accuracy is everything right over everything: (80 + 890) ÷ 1,000 = 97.0%. Precision is how many flagged cases were real: 80 ÷ 100 = 80.0%. Recall is how many real frauds were caught: 80 ÷ 90 = 88.9%. Specificity is how many legitimate transactions were passed: 890 ÷ 910 = 97.8%. F1 is the harmonic mean of precision and recall: 2 × 0.800 × 0.889 ÷ (0.800 + 0.889) = 84.2%. Accuracy looks best only because fraud is rare, which is why it is a poor guide on unbalanced data.
Question 3 · Matching · Machine learning
Five business problems have reached the data science team, and each suits a different kind of algorithm.
Match each problem to the algorithm that fits it.
Linear regression predicts a number from features. Logistic regression predicts the probability of one of two classes. K-means groups unlabelled records into clusters. Principal component analysis reduces many correlated features to a few components that keep most of the variation. ARIMA forecasts a series from its own past values. Apriori finds items that are often bought together, and Q-learning is a reinforcement learning method; neither fits these problems.
Question 4 · Ordering · Operations and processes
A data science project is being planned with CRISP-DM, the cross-industry standard process for data mining, and the project board wants to see the phases in order.
Arrange the six phases from first to last.
The process starts with the business question, because a model that answers the wrong question is worthless. Then comes understanding the data that exists, and preparing it, which usually takes most of the time. Modeling and evaluation follow, where evaluation asks whether the model meets the business goal and not only whether it scores well. Deployment puts it to use. In practice the phases loop back: evaluation often sends a project back to business understanding.
Question 5 · Matching · Specialized applications of data science
Five product teams described what they need, each in one sentence, and each need matches an established technique.
Match each need to the technique.
Tokenization splits text into the units a language model works with. Named entity recognition tags the people, places and organizations mentioned in text. Object detection finds items in an image and draws a box around each. Collaborative filtering recommends what similar users liked. Anomaly detection flags records that are unlike the rest, which is the basis of most fraud screening. Backpropagation is how neural networks are trained, and one-hot encoding turns categories into numbers.
CompTIA lists DataAI among its certifications with simulation performance-based questions: tasks in an approximation of a real tool, such as a firewall, a network diagram or a terminal window. CompTIA’s page on performance-based questions, read on November 12, 2026.
Guide
CompTIA describes performance-based questions as exam items that test your ability to solve problems in real-world settings. Most are simulations: an approximation of a tool such as a firewall, a network diagram, a terminal window or an operating system.
You don’t pick an answer. You do the task, and there can be more than one correct way to do it.
Try each question cold. Getting it wrong first and then reading why teaches more than reading the answer and nodding along.
A right answer reached by guessing is a wrong answer waiting for exam day. Every question here explains why each part is what it is.
CompTIA says a simulation question can be skipped and returned to later, with your work saved. If one is eating your time, move on, answer the multiple-choice questions and come back.
14 free labs in your browser: Linux and Windows command line, firewall rules and SOHO router setup, checked as you go.
Open the toolFree performance-based questions for 19 CompTIA exams: firewall rules, network diagrams, command-line tasks and drag-and-drop.
Open the toolFree practice tests for 261 popular certification exams, with explanations and a score by exam domain.
Open the toolFAQs
Find quick answers or contact our support team.
Yes. CompTIA lists DataAI among its certifications with simulation performance-based questions: tasks in an approximation of a real tool, such as a firewall, a network diagram or a terminal window. That is from CompTIA’s page on performance-based questions, read on November 12, 2026.
No. They’re original tasks written to each exam’s published objectives. Using leaked real exam questions breaks every vendor’s exam rules.
CompTIA says a performance-based question can have more than one correct way through it, that its scoring allows for the different approaches, and that partial credit may be given. Here, each question is scored by its parts, so you can see which part you missed.
CompTIA says a simulation question can be skipped and returned to later, and that your work on it is saved when you move on. A virtual-environment question has to be completed when you reach it: a warning screen tells you so first.
No. The questions run in your browser and nothing is saved. Sign up only if you want full practice exams and progress tracking.
OpenpassAI
Hi 👋 I’m OpenpassAI. Ready to prepare for your exam the smarter way?
Ready to explore OpenpassAI? I can show you how the platform helps you prepare, practice, and pass.
