A data analyst assessment test is a timed screening exercise that asks you to write SQL, work a spreadsheet, answer quiz questions on statistics and explain what a dataset means for the business. Published formats run from 10 minutes of multiple choice to 90 minutes of timed modules, and the test usually lands before your first conversation with a human.
What the test usually contains
Expect SQL first. Candidates report joins, WHERE filters, GROUP BY aggregates and window functions such as RANK or LAG. Codility runs SQL tasks against a live PostgreSQL or SQLite database and scores the query output, not your syntax.
Spreadsheets come next. Excel or Google Sheets tasks ask for pivot tables, VLOOKUP or INDEX MATCH and a basic chart. CodeSignal hands you CSV files to download and lets you use Excel, Google Sheets, R or Python.
Then statistics. Quiz items ask you to read a chart without being fooled by an average, pick the right metric and explain a trend in plain language. Codility's knowledge sets are 10-question multiple-choice blocks.
The last section is usually a case. You get a business problem, sometimes a small dataset, and are asked what you would recommend. Tech take-home versions give 4 to 7 days, then a 45 to 60 minute presentation.
Formats and platforms that host it
| Platform | Format | Time |
|---|---|---|
| CodeSignal Data Analytics | 1 coding question, 14 quiz questions, CSV files | 70 minutes |
| Codility SQL | Written queries against a live database | 30 to 35 minutes per task |
| TestGorilla Working With Data | Questions on data handling, analysis and charts | 10 minutes |
| TestGorilla Data Analyst interview | 8 interview questions | 26 minutes |
| CodeSignal Data Science Frameworks | 5 modules, scores from 200 to 600 | 90 minutes |
| HackerRank Data Analyst | Role-based test from its certified library | Set by the employer |
HackerRank's Data Analyst role page lists SQL at advanced level, plus R, Python, statistics, Excel, Power BI and Tableau.
How to prepare
Practise on a clock. Write joins, aggregations and window functions by hand with no autocomplete, then build a pivot table and a lookup in a blank workbook. Speed decides it once the skill is there.
Check the tool rules in your invite before the timer starts. If the platform allows Python, confirm that pandas is available. Candidates report that many Python sections ban pandas, NumPy and SciPy and expect plain functions over lists and dictionaries.
Do two full runs on the platform's own practice material, timed end to end.
What employers should check before sending one
Match the task to the job. Codility's guidance is that a data analyst assessment can use SQL tasks only, while a data engineer assessment can combine SQL with a Python task. Sending Python to an Excel-heavy role measures the wrong skill.
Set the cut score before anyone sits the test. Codility lets you set it, and CodeSignal reports a certified score with no published pass mark. For a version with no take-home at all, see how to test candidate skills without a take home assignment.
Scoring is where programs drift. An AI Candidate Evaluation run against the same rubric for every candidate gives a comparison a hiring manager can defend, and it pairs with a short note to anyone who did not pass, using this follow-up email to a candidate after failed assessment. If scores cluster near the top, read how to tell if a candidate cheated on a coding assessment.