🚧 Under construction — VoTechWorks is being built. A free tech program — complimentary, as our commitment to get America working.
voTechWorksGrade 13 · a vocational technology learning center, by urbanicity
newSchoolWorks.com › VoTechWorks › Technology & Design › Data Science & Analytics › Certification Exam Preparation: Google Data Analytics, PL-300, IBM Data Science, and CompTIA Data+ › Lesson 3
TEC-DATA85-14 · Lesson 3 · 24 clock hours · lab

Domain-by-Domain Targeted Review: Microsoft PL-300 (Power BI) and IBM Data Science Professional

Objective. For each domain listed in the Microsoft PL-300 exam skills measured document and the IBM Data Science Professional Certificate objectives, the learner will complete domain-focused practice question sets and hands-on skill tasks, achieving at least 75 percent accuracy per domain.

⚠ Pending professional review

This lesson, its worked examples and its practice questions has not yet been verified by an industry-certified professional or by a certification or approval body. It is a study draft and may contain errors. Do not use it as instructions for work on a job site or on live equipment. Always follow your instructor, your supervising professional, and current official references and regulations.

VoTechWorks program approval and accreditation are still in progress. Credentials are issued only by the certifying bodies named in each program, after their own exams.

Domain-by-Domain Targeted Review: Microsoft PL-300 and IBM Data Science Professional

Targeted domain review is the most efficient way to close knowledge gaps before a certification exam. Rather than re-reading every chapter, you identify which skill areas are below threshold, drill those areas with timed practice questions, and then verify understanding by doing — building a measure in Power BI Desktop or running code in a Jupyter Notebook. This lesson walks through that process for both the Microsoft PL-300 (Power BI) exam and the IBM Data Science Professional Certificate.

How Targeted Review Works

Both credentials publish their measured skill areas openly. Microsoft posts the current PL-300 Exam Skills Measured document on Microsoft Learn; IBM structures its certificate around defined course modules on Coursera. Before you open a single practice question, download or bookmark the skills document for PL-300. Read every bullet. Each bullet is a potential exam item. Do the same for the IBM program's stated learning objectives. These documents are your syllabus for this lab.

Once you have the domain list, work one domain at a time. Complete the 20-question set for that domain, score it, and log the result. If you score 75 percent or above, move to the next domain. If you fall below 75 percent, stop, identify which specific questions you missed, and address each gap — either by reading the relevant Microsoft Learn module, reviewing the IBM course material, or, most powerfully, reproducing the scenario in live software.

PL-300 Skill Areas in Detail

The PL-300 exam organizes its content into four broad skill areas: preparing data in Power BI, modeling data, visualizing and analyzing data, and deploying and maintaining assets. Weights shift slightly as Microsoft updates the exam, so always check the current Skills Measured document.

Preparing data covers connecting to data sources, profiling data for quality issues, and applying transformations in Power Query Editor. Common exam traps include the difference between removing duplicates at the query level versus the visual level, and understanding how query folding affects performance with relational sources.

Modeling data is typically the heaviest domain. You must understand the star schema — fact tables, dimension tables, and why many-to-many relationships require careful handling. DAX (Data Analysis Expressions) is tested here. Know the difference between a calculated column and a measure: a calculated column stores a value row-by-row in the table; a measure is computed at query time in the context of a visual. Practice writing CALCULATE, FILTER, ALL, and time-intelligence functions such as TOTALYTD. When a practice question about DAX trips you up, open Power BI Desktop, load a small dataset, write the measure, and check the result in a card visual. Seeing the correct output locks the concept in memory far better than re-reading the definition.

Visualizing and analyzing data covers chart selection, conditional formatting, report design accessibility, Q&A visuals, AI visuals such as Key Influencers, and using slicers and filters effectively. Exam questions in this domain often present a business requirement and ask which visual type or configuration best meets it.

Deploying and maintaining assets covers publishing to the Power BI service, configuring workspaces and apps, setting up scheduled refresh, managing row-level security (RLS), and understanding sensitivity labels. These topics are sometimes under-studied — give them full attention.

IBM Data Science Professional Domain Areas

The IBM certificate is a multi-course program. For exam and capstone preparation, focus on these core competency areas:

Data science methodology — understanding the structured process of moving from business question to data collection to model deployment. Know the phases and be able to apply them to a scenario.

Python for data science — data types, Pandas DataFrames, data cleaning with `.dropna()` and `.fillna()`, and basic statistical summaries. If a practice question about Pandas syntax is missed, open a Jupyter Notebook and run the correct syntax immediately.

Data visualization — using Matplotlib and Seaborn to build charts that communicate findings. Know when to use a bar chart versus a histogram versus a scatter plot.

Machine learning with Scikit-learn — the standard workflow: split data with `train_test_split`, instantiate a model, call `.fit()`, call `.predict()`, and evaluate with metrics. For classification problems, understand accuracy, precision, recall, and the confusion matrix. A confusion matrix is a table showing true positives, true negatives, false positives, and false negatives. Logistic regression predicts a categorical outcome despite the name containing "regression."

Applied capstone competencies — integrating all prior skills: loading a CSV, cleaning it, performing exploratory analysis, building a model, and communicating results.

The Integrated Lab: Connecting Both Worlds

The 90-minute integrated lab in this lesson ties PL-300 and IBM skills together. You will load a raw CSV into Power BI Desktop and a Jupyter Notebook simultaneously. In Power BI, build a clean data model and write at least one DAX measure. In Python, clean the same dataset, train a logistic regression model, and display a confusion matrix. Then answer 10 scenario-based questions referencing the specific steps you just completed. This mirrors the applied, scenario-based item style that both credentials favor.

Accuracy, Integrity, and Workplace Safety Considerations

Data work carries professional responsibility. Reporting incorrect metrics because a DAX measure was written carelessly, or misclassifying outcomes because a model was trained on unscaled data, can lead to poor business decisions. Treat accuracy during practice as you would on the job. In professional settings, data pipelines that connect to live systems must be configured with appropriate credentials and access controls — never store passwords in plain text in a notebook or a Power BI connection string. Sensitivity labels in the Power BI service exist precisely to protect data that is confidential or regulated. When working on shared lab machines, log out of all cloud services when the session ends to protect account credentials.

Tracking Your Progress

Keep a domain score log for every practice session. A simple table with domain name, date, score, and notes on missed topics is sufficient. Share it with your instructor at check-in. The threshold is 75 percent on every domain — not just an average across all domains. A strong score in one area does not excuse a weak score in another.

Key terms8

DAX (Data Analysis Expressions)A formula language used in Power BI, Analysis Services, and Excel Power Pivot to create calculated columns, measures, and tables; measures are evaluated in the context of the visual they appear in rather than stored row-by-row.
Measure (Power BI)A DAX calculation that is computed dynamically at query time based on filter context from slicers, rows, columns, and report filters; unlike a calculated column, a measure does not add a physical column to the table.
Star SchemaA data modeling pattern where a central fact table containing numerical, measurable data connects to surrounding dimension tables containing descriptive attributes; the preferred structure for Power BI data models.
Row-Level Security (RLS)A Power BI feature that restricts which rows of data a given user can see when viewing a report or dashboard, enforced through DAX filter rules assigned to security roles.
Confusion MatrixAn evaluation table for classification models that shows the counts of true positives, true negatives, false positives, and false negatives, allowing calculation of accuracy, precision, and recall.
Logistic RegressionA supervised machine learning algorithm used for binary or multiclass classification that estimates the probability of class membership; despite the name, it predicts categorical outcomes.
Query FoldingA Power Query behavior in which transformation steps are translated into a native query (such as SQL) and executed on the source system rather than in Power BI, improving performance with compatible data sources.
train_test_splitA Scikit-learn utility function that randomly divides a dataset into separate training and testing subsets so that model performance can be evaluated on data the model has not seen during training.

Worked examples3

1. A logistic regression model trained on a binary classification dataset produces the following confusion matrix results: True Positives = 40, True Negatives = 35, False Positives = 10, False Negatives = 15. Calculate accuracy, precision, and recall.
  1. Identify the four values: TP = 40, TN = 35, FP = 10, FN = 15. Total predictions = 40 + 35 + 10 + 15 = 100.
  2. Calculate accuracy: (TP + TN) / Total = (40 + 35) / 100 = 75 / 100 = 0.75, or 75 percent.
  3. Calculate precision: TP / (TP + FP) = 40 / (40 + 10) = 40 / 50 = 0.80, or 80 percent.
  4. Calculate recall: TP / (TP + FN) = 40 / (40 + 15) = 40 / 55 ≈ 0.727, or approximately 72.7 percent.
Answer: Accuracy = 75%, Precision = 80%, Recall ≈ 72.7%
2. In Power BI Desktop, a Sales table has a Revenue column. Write a DAX measure that calculates total revenue only for the current year using a time-intelligence function, and explain each part.
  1. In the Data pane, right-click the Sales table and select 'New measure.'
  2. Write the measure: Revenue YTD = TOTALYTD(SUM(Sales[Revenue]), 'Date'[Date])
  3. TOTALYTD is a time-intelligence function that accumulates a value from the start of the year to the current date context. SUM(Sales[Revenue]) is the expression being accumulated. 'Date'[Date] is the date column from a properly marked Date table.
  4. Verify: place the measure in a table visual alongside a Year and Month column. Revenue YTD for March should equal the sum of January + February + March revenue for that year.
Answer: Revenue YTD = TOTALYTD(SUM(Sales[Revenue]), 'Date'[Date])
3. A Power BI data model has a Sales fact table and a Products dimension table. A many-to-many relationship exists between them through a shared Category column (not a unique key). Calculate the row count that a measure would see if the Products table has 3 category rows each repeated 2 times (6 rows total) and Sales has 12 rows, and explain why this relationship requires a bridge table or a corrected model.
  1. Count distinct categories in Products: the column has 3 unique values but 6 physical rows because each category appears twice — this violates the dimension table rule that the join key must be unique on the 'one' side.
  2. In a many-to-many relationship without a bridge table, Power BI cannot guarantee filter propagation direction; a filter on one category in Products may match multiple rows, causing measures to double-count or behave unexpectedly across the 12 Sales rows.
  3. The correct fix: ensure the Products dimension table has one row per unique Category (3 rows), giving a proper one-to-many relationship from Products[Category] to Sales[Category].
  4. After correction, a CALCULATE measure filtered to one category will correctly scope to the subset of the 12 Sales rows that belong to that category — no duplication.
Answer: The many-to-many relationship causes ambiguous filter context; the dimension key must be unique. Correct the Products table to 3 unique category rows to establish a clean one-to-many relationship.

The math in this lesson1

Learners compute accuracy, precision and recall from a confusion matrix for exam review.
S-CP.4
Brush up in the math halls at newSchoolWorks.com:
In class: Work through the Microsoft PL-300 exam skill areas — including preparing data in Power BI, modeling data, visualizing and analyzing data, and deploying and maintaining assets — by completing 20-question practice sets per skill area using Microsoft Learn's official PL-300 practice assessment and third-party practice banks; for modeling and DAX questions, open Power BI Desktop, reproduce the scenario in a live dataset, and verify the correct measure or relationship before marking the answer. · Work through the IBM Data Science Professional Certificate domain areas — including data science methodology, Python for data science, data visualization, machine learning with Scikit-learn, and applied data science capstone competencies — by completing 20-question practice sets using Coursera quiz banks from the IBM program and supplementary flashcard decks; for any machine learning question answered incorrectly, open a Jupyter Notebook and implement a minimal working example of the algorithm or evaluation metric in question to reinforce conceptual understanding with code. · Complete a 90-minute integrated lab: given a raw CSV dataset, load it into both Power BI Desktop and a Jupyter Notebook, build a clean data model and one DAX measure in Power BI, then build a logistic regression model and produce a confusion matrix in Python — then answer 10 examiner-style questions drawn from PL-300 and IBM domains that reference specific steps taken in the lab, simulating the applied and scenario-based item types used in both credentials.
Materials: Microsoft PL-300 Exam Skills Measured document (current version, free from Microsoft Learn); Microsoft Learn PL-300 practice assessment; Power BI Desktop (free download); IBM Data Science Professional Certificate course materials on Coursera; Jupyter Notebook (Anaconda distribution); Scikit-learn documentation; sample CSV datasets from course 5 and course 8; timer.
Check: Instructor reviews domain score logs and the integrated lab write-up; learner must reach at least 75 percent on every PL-300 skill area and every IBM domain set, and the integrated lab must include a functioning Power BI file and a Jupyter Notebook that executes without errors — pass requires all domains at threshold and a complete, working lab submission.

Practice check8

Answer each question, then check it. Nothing is recorded; this is your own practice.

Microsoft PL-300 — Modeling Data
1. In Power BI Desktop, you need a value that is computed dynamically based on the current slicer selection and does not add a column to the underlying table. Which object type should you create?
Microsoft PL-300 — Deploying and Maintaining Assets
2. A report user should only see sales data for their own region when they view a Power BI report published to the Power BI service. Which feature enforces this restriction?
Microsoft PL-300 — Preparing Data in Power BI
3. When applying transformations in Power Query Editor against a SQL Server source, which behavior should you look for to confirm that the transformations are being processed on the database server rather than in Power BI?
Microsoft PL-300 — Modeling Data
4. A TOTALYTD DAX measure is returning blank values for some months in a table visual. What is the most likely cause?
IBM Data Science Professional — Data Science Methodology
5. In the IBM Data Science methodology, at which phase does the practitioner first define the analytic approach — for example, deciding whether to use a classification or clustering model?
IBM Data Science Professional — Machine Learning with Scikit-learn
6. A logistic regression model for predicting customer churn shows 90 percent accuracy but very low recall on the churned class. What does this most likely indicate?
IBM Data Science Professional — Machine Learning with Scikit-learn
7. You run a logistic regression model and obtain: TP = 50, TN = 30, FP = 20, FN = 0. What is the model's recall?
Microsoft PL-300 — Visualizing and Analyzing Data
8. In Power BI Desktop, a visual-level filter is applied to a bar chart to exclude a specific product. A page-level filter that includes all products is also active. Which filter takes precedence for that visual?
Record this lesson on my transcript. Sign in with your key and enrollment email to record lessons (enroll free if you have no key). This page stays open to read and practise.
← Previous lessonProgram planNext lesson →