Updated November 2026
TL;DR
A data analyst's job breaks into six parts: turning a vague question into one data can answer, pulling the data out of a database, cleaning and checking it, running the comparison that answers the question, building a report or dashboard, and explaining the result to people who will never open the file. Most entry-level analyst roles mix these six, usually with SQL, a spreadsheet and a dashboard tool. The best test of whether you'd like the work is to answer one question end to end with a free public dataset this week and notice which parts you enjoyed and which you dreaded.
Job postings describe the role as "extracting actionable insights," which tells a high schooler almost nothing. What sets this guide apart is that it describes the job in terms of what an analyst is actually typing, checking or arguing about, with one thing you can try for each part.
Want to try it for real? See Ladder's lists of data analyst internships and data science internships for high school students.
What do you need to know about data analyst jobs?
The Bureau of Labor Statistics doesn't track "data analyst" as its own occupation; it counts much of the work under data scientists.
|
Detail |
What you need to know |
|
Core tools |
SQL, spreadsheets and a dashboard tool (Tableau, Power BI or Looker Studio) |
|
Typical education |
Bachelor's degree, often in math, statistics, computer science, business or engineering |
|
Median pay (data scientists, May 2025) |
$120,230 a year; lowest 10% under $67,240 |
|
Projected growth (data scientists) |
35% from 2025 to 2035, much faster than average |
|
Most time-consuming part |
Cleaning and checking data |
|
How to try it now |
Free tutorials, public datasets and free dashboard tools |
Why should you understand what a data analyst does before choosing it?
Because "analyst" covers jobs that barely overlap: one posting wants someone writing database queries all day, another wants someone rebuilding a weekly report and answering questions about it, and a third is closer to finance. There's no single job behind the title.
That matters because you'll make real choices from your picture of the work, like whether to take statistics, which clubs to join and which internships to apply for. Job-posting language makes the role sound strategic and leaves out the hours spent checking whether a number is right.
What does a data analyst's week look like?
Busy and varied: standing meetings, quick questions by message and longer stretches of building something properly. Work arrives from several directions, which is a big part of why analysts end up understanding how a whole business fits together.
A five-minute question can turn into a real investigation once two reports disagree, and that detour is often where the interesting finding is. The six parts below show what each piece of the job looks like and how to try it now.
The 6 parts of a data analyst's job
1. A data analyst turns a vague request into an answerable question
"Why were sales down last month?" isn't a question data can answer yet: down compared to what, in which region, in orders or in revenue? Analysts who skip this step deliver a correct number that answers a question nobody asked.
Try it now: take a fuzzy question from your life, such as whether your debate team is improving, and rewrite it as something a spreadsheet could settle, with the exact comparison and time period. Do it ten times.
2. A data analyst pulls the data out of a database
The data usually lives in a database, and you get it out with SQL, a language for asking for specific rows and columns: for example, every order from the last 12 months, joined to the customer table, grouped by region, excluding refunds. A query that runs isn't always a query that's right.
Try it now: free interactive tutorials such as SQLBolt and Mode's SQL tutorial let you write real queries in your browser, and Kaggle has free datasets and short SQL courses. Aim for filtering, sorting, grouping and joining two tables.
3. A data analyst cleans the data and checks whether it's trustworthy
This is the part that eats the most time: the same city spelled four ways, dates in two formats, duplicate rows and blank fields. The BLS describes it as structuring raw data so software can read it. The harder question underneath: if two reports give different totals for the same month, which one is wrong?
Try it now: download a messy file from a city or government open data portal, Our World in Data or Kaggle, and clean it in Google Sheets until every column is consistent and you can defend every row you deleted. Keep notes on what you changed.
4. A data analyst runs the comparison that answers the question
Most entry-level analysis is comparing, counting and averaging, not machine learning: customers who used the new checkout against those who didn't, or this quarter against the same quarter last year. The key skill is knowing when a gap is big enough to act on and when it's noise, which is where statistics pays off.
Try it now: pick two groups in a dataset you've cleaned, decide in advance what result would change your mind, then check whether the difference survives the messiness you found. Or find a claim in a news article and check it against the public data.
5. A data analyst builds the dashboard people keep coming back to
A one-off answer gets asked for again next month, so analysts build dashboards: pages of charts wired to live data that a manager can open instead of messaging you. Maintaining them matters as much as building them, because a dashboard breaks quietly when someone renames a column.
Try it now: Tableau Public and Looker Studio are free, and you can connect either to a Google Sheet and have a working dashboard within an hour. Build one you'd actually check, like your team's race times, then watch which chart someone else misreads first.
6. A data analyst explains the result to people who won't open the file
A finding only counts once someone acts on it, so analysts spend real time in meetings and documents saying in three sentences what took three days. The BLS notes that data scientists communicate findings to both technical and non-technical audiences, and the hardest part is answering honestly when someone asks how confident you are.
Try it now: compress an analysis into one slide and 90 seconds (the question, the answer, the one number that carries it and what you'd do) and present it to a club or a teacher who'll push back.
How can you build data analyst skills in high school?
Stack the six try-it-now ideas into one project: pick a dataset you care about, sharpen a question, pull and clean the data, run the comparison, build a small dashboard and write up what you found in under 500 words. A finished project proves you can handle the unglamorous middle of the work, and posting it publicly lets anyone see your skills in minutes.
What a solo project can't give you is someone else's question and a person with the authority to push back on your answer. Ladder Internships places high school students in remote, eight-week start-up internships, where they work on a real project with a company coach, including in data and analytics, and finish by presenting to the team, with full financial aid available.
Common questions about what a data analyst does
Will AI replace data analysts?
No one can promise either way, but the BLS projects data scientist jobs to grow 35% from 2025 to 2035, partly because companies are adding AI. Routine queries and standard charts are most exposed; knowing which question to ask and whether a number can be trusted is harder to automate.
Do you need a computer science degree to be a data analyst?
No. Data scientists typically have a bachelor's in math, statistics, computer science, business or engineering, and analysts also come from economics, psychology or public health. In high school, statistics and strong algebra will help most.
How much do data analysts make?
The BLS doesn't publish a separate data analyst wage. For data scientists, the median was $120,230 in May 2025, with the lowest 10% under $67,240; entry-level analyst pay is usually well below that median and varies by city and industry.
How can you try data analysis on a real project?
Work with a real team that's waiting on your answer. Ladder Internships runs remote, eight-week start-up internships for high school students worldwide, at 5 to 10 hours a week, with a company coach and full financial aid available. The next deadline is November 15, 2026, for the Winter I cohort starting December 7, 2026.
Contact us at contact@ladderinternships.com if you have any questions!