Data Analytics Test

Data Analytics Test

Data analytics interviews tend to circle the same ground. A SQL join question. Something about mean versus median. Whether you know that correlation is not causation, and whether you can explain why.

This free data analytics test covers 20 of those fundamentals across SQL, statistics, spreadsheets, charts, Python, data cleaning, A/B testing, and the analysis process itself. Every question shows you the right answer and a short explanation the moment you choose, so a wrong answer teaches you something instead of just costing you a point.

You get a score out of 100, a readiness level, and a certificate you can download for a portfolio or LinkedIn. It takes about eight minutes. No sign up, no email, and nothing you answer leaves your browser.

How the Data Analytics Test Is Scored

Twenty questions, five points each, and the answer options shuffle every time so you cannot memorize positions.

Correct answersScoreLevel
18 to 2090 to 100Interview Ready
15 to 1775 to 85Strong Candidate
12 to 1460 to 70Solid Foundation
8 to 1140 to 55Building Skills
7 or fewerBelow 40Just Starting

Worth being clear about what a high score does and does not mean. Recognizing the right definition is the easy half. Real interviews ask you to write a query, read a messy chart, or explain your reasoning out loud, and those are separate skills built by practicing on actual data.

What the Test Covers

AreaWhat comes up
SQLJoins, GROUP BY, primary keys, and what the language is for
StatisticsMean, median, mode, outliers, correlation and causation
VisualizationPicking the right chart, histograms, dashboards
SpreadsheetsLookup functions and pivot tables
PythonWhich library does what
ProcessData cleaning, A/B testing, KPIs, types of analytics, where a project starts

SQL: The Part Interviews Lean On Hardest

If you only study one area, make it this one. SQL comes up in almost every analytics interview, and it is usually where candidates get filtered out.

Joins

A join combines rows from two tables using a column they share, like matching orders to customers by customer id. Which join you use decides what happens to rows that have no match.

Join typeWhat you get back
INNER JOINOnly rows that have a match in both tables
LEFT JOINEvery row from the left table, plus matches from the right. Unmatched right side comes back empty.
RIGHT JOINThe mirror image, every row from the right table
FULL OUTER JOINEvery row from both, with gaps where there was no match

The classic interview trap is asking which join finds customers who have never placed an order. The answer is a LEFT JOIN from customers to orders, keeping only the rows where the order side came back empty. An INNER JOIN would silently drop exactly the people you were looking for.

GROUP BY and aggregates

GROUP BY collapses rows that share a value so you can total or count them by category. Revenue per country, orders per customer, average score per class. It works alongside aggregate functions such as SUM, COUNT, AVG, MIN, and MAX.

One distinction interviewers like: WHERE filters individual rows before grouping, while HAVING filters the groups after. Finding countries with more than a thousand orders needs HAVING, because that total does not exist until the grouping has happened.

Primary keys

A primary key is the column that uniquely identifies each row, like a customer id. It is what makes joins reliable, because without a unique identifier you cannot be certain which row is which.

The Statistics That Actually Come Up

Mean, median, and mode

The mean is the average, the median is the middle value once everything is sorted, and the mode is whichever value appears most often.

The test uses the set 2, 3, 3, 6, 100. The median is 3, because it sits in the middle. The mean is 22.8, because one enormous value drags it upward. Nothing in that data is anywhere near 22.8, which is exactly the point.

Why this matters outside the classroom

This is why average salary and average house price can be so misleading. A handful of very high values pull the mean far above what a typical person experiences. When a distribution is lopsided, the median usually describes reality better, and knowing when to reach for it is a genuine analytical skill rather than a definition to recite.

Outliers

Outliers sit far from everything else. The instinct is to delete them, and that instinct is often wrong. Sometimes they are data entry errors worth removing. Sometimes they are the most important thing in the dataset, like the fraud case or the one product that took off. Investigate before you discard.

Correlation and causation

Two things moving together does not prove one causes the other. Ice cream sales and drownings both rise in summer, and neither causes the other. Heat drives both.

Three explanations are worth having ready for an interview.

  • A hidden third factor. Something else drives both, as with the heat example.
  • Reverse causation. The relationship runs the other way. Cities with more police also have more crime, but the crime came first.
  • Pure coincidence. Search enough datasets and you will find things that track each other perfectly for no reason at all.

Choosing the Right Chart

What you want to showUse
Change over timeLine chart
Comparing categoriesBar or column chart
How one numeric variable is spreadHistogram
Relationship between two numeric variablesScatter plot
Parts of a whole, with very few slicesPie chart, used sparingly
Values across geographyMap

Pie charts get criticized for a reason. People judge angles poorly, so once you have more than about four slices a bar chart communicates the same thing far more clearly.

How charts mislead

Being able to name these is genuinely useful, and it comes up more often than people expect.

  • A truncated axis. Starting the vertical axis at 95 rather than zero turns a tiny change into a cliff.
  • A cherry picked date range. Choosing exactly the window where the trend supports your point.
  • Dual axes. Two different scales side by side can be positioned to make unrelated lines appear to move together.
  • Percentages with no base. A 200 percent increase means something very different from two cases to six than from two million to six million.
  • 3D effects. Depth distorts the proportions people are trying to read.

Spreadsheets and Python

Most analytics work still happens in spreadsheets, and interviewers know it.

Lookup functions find a value in one column and return the matching value from another, like getting a price from a product id. VLOOKUP is the traditional one and XLOOKUP is the newer replacement, which is more flexible and does not break when columns move.

Pivot tables group and summarize without any formulas at all. Sales by region by month in a few clicks. If you can build one confidently you can answer a large share of everyday business questions.

On the Python side, knowing which library does what is usually enough for an early interview.

LibraryWhat it is for
pandasWorking with tables of data. The DataFrame is a spreadsheet you control with code.
NumPyFast numerical arrays, sitting underneath much of pandas
Matplotlib and seabornDrawing charts
scikit-learnMachine learning models

The Four Types of Analytics

TypeThe question it answersExample
DescriptiveWhat happened?Sales fell 12 percent last quarter
DiagnosticWhy did it happen?The drop came from one region after a price change
PredictiveWhat is likely next?These customers are at risk of leaving
PrescriptiveWhat should we do?Offer this segment a retention discount

Most day to day analyst work sits in the first two. The later two get the attention, but a business that cannot answer what happened and why has no foundation for anything else.

Data Cleaning and A/B Testing

Cleaning takes up a large share of the job, and it is the least glamorous part of every analytics role. Duplicate rows, missing values, dates stored as text, inconsistent spellings of the same category, numbers with stray currency symbols. Every one of those quietly breaks an analysis if it survives into the results.

A/B testing splits people between two versions and measures which performs better on one chosen metric. The parts that trip candidates up are the discipline around it rather than the concept. You need enough traffic for the result to mean anything, you should decide the success metric before you start, and stopping the moment the numbers look good is a reliable way to fool yourself, because random noise will hand you a favorable moment eventually if you keep watching.

How to Prepare for a Data Analyst Interview

  1. Write real SQL, not just read it. Free practice sites and any public dataset will do. Recognizing a join is not the same as writing one under pressure.
  2. Have one project you can walk through end to end. The question, where the data came from, what you cleaned, what you found, what you would do differently. This carries more weight than another certificate.
  3. Practice explaining out loud. A lot of the interview is whether a non technical manager would understand you.
  4. Know your definitions cold. Mean versus median, correlation versus causation, the join types. These are free points and getting one wrong stands out.
  5. Prepare for the messy question. Something like why did sign ups drop last month. They want to see how you break a problem down, not a memorized answer.
  6. Be willing to say you do not know. Followed by how you would find out. That answer lands better than a confident guess, in the interview and in the job.
This is a free practice test for learning and self assessment. It is not a professional certification, not affiliated with any employer or certifying body, and the certificate is a keepsake rather than a credential. Use it to find your gaps and guide your study.

Frequently Asked Questions

What is a good score on this data analytics test?

Anything at 90 or above suggests the fundamentals are solid. Between 75 and 89 means you know most of what an interviewer will probe with a few gaps to close. Below 60 points to specific areas worth studying, and the explanations shown during the test tell you which.

What skills does a data analyst need?

SQL first, then spreadsheets, then a grasp of basic statistics and how to choose a chart. Python or R helps and matters more in some roles than others. Underneath all of it sits the ability to define a clear question and explain your answer to someone non technical.

What is the difference between mean and median?

The mean is the average of every value. The median is the middle value once they are sorted. Extreme values pull the mean around but barely move the median, which is why the median describes skewed data like income or house prices more honestly.

What is the difference between an INNER JOIN and a LEFT JOIN?

An INNER JOIN returns only rows that match in both tables. A LEFT JOIN returns every row from the left table whether or not it matched, filling the missing side with empty values. Use LEFT when you need to keep the unmatched rows, such as finding customers with no orders.

Do I need Python to be a data analyst?

Not for every role. Plenty of analyst jobs run on SQL and spreadsheets alone. Python widens your options and becomes more important as roles move toward larger datasets and automation.

How long does this test take?

Around eight minutes for 20 questions. There is no timer, and each answer is explained as you go.

Can I retake it?

As often as you like. Options reshuffle each time, so remembering positions will not help you.

Is it free, and are my answers stored?

Free, with no sign up or email. Everything runs in your browser and nothing you select is saved or sent anywhere.

More Free Tests

  • Data Analytics TestYou are here. Twenty questions on SQL, statistics, charts, and process, with explanations and a certificate.
  • Critical Thinking ExerciseTwelve scenarios that train you to spot weak reasoning and misleading claims.
  • AI Literacy TestHow AI really works, including where it fails and why output needs checking.
  • Marketing IQ TestThe metrics side of marketing, from conversion rate to A/B testing.

Other Tools You Might Like