Data Analyst
You’re the person who finds the story hidden in a pile of numbers, the “wait, that’s why signups dropped” moment when the data finally clicks. The tools can write the SQL (the language for pulling data out of a database) and build the charts now, so the part that’s really yours is knowing which question is even worth asking.
Related: Quantitative Analyst, Journalist, Product Manager
The day in the life
Perch · new accounts / week
last 6 weeks · down ~30%
Fewer people are making an account. Where do you look first?
Try a day as a data analyst
A short, playful taste of the real work.
Lots of work for people who can find the story in a pile of numbers, but the simple tasks you would start on are the ones computers now do by themselves.
No strict degree rule, but you need to be good at asking a company’s stored records questions, using a language built for that called SQL.
No standalone BLS code; closest is Operations Research Analysts (median ~$90K, May 2024); Data Scientists (15-2051, median $112,590) for the higher tier; Glassdoor / Levels.fyi 2026.
What you’d actually do
The picture is cool insights and storytelling. The reality is mostly cleaning messy data, writing SQL (the language for pulling data out of a database), and rebuilding the same reports each week for whoever asks.
The tools can now draft those queries and build the dashboards. So the part that stays yours is judgment: asking the right question, checking a draft didn’t quietly get it wrong, and turning the result into a decision someone will act on.
- Data gathering & cleaning50%
- Analysis & modeling15%
- Dashboards & charts20%
- Explaining what it means5%
- Meetings & admin10%
Half the day is tidying up messy numbers so they can be used at all, plus building charts. Nobody asks you what it all means yet.
Rough split, based on how analysts describe the work. Varies by company and tools.
The tools stay similar; what you’re analyzing changes everything else.
- Business analyticsfiguring out what’s happening inside a company right now.
- Marketing analyticsfiguring out which campaigns and channels are actually working.
- Product analyticsfiguring out how people actually use a piece of software.
- Financial analyticsfiguring out where the money’s going and what it means.
A typical early-career day
- 9:30Take the request
A stakeholder asks "can you pull X." Figure out what they actually need, which is half the work.
- 10:30Pull & clean the data
Get the data and wrestle it into shape. Messy, missing, and inconsistent is the normal starting point.
- 1:00Write the query
Write the SQL or analysis to actually answer the question. A tool can draft this now, but you steer it.
- 3:00Check it’s really right
Sanity-check the result. A query can run fine and still be quietly, confidently wrong.
- 4:30Draft fast, you judge
The queries and dashboards get drafted fast; your value is the question it can’t pick and the error it can’t catch.
A rough analyst day. The repetitive parts are real, and as those parts get automated, the job shifts toward framing questions and judging answers.
The outlook
Where it’s going
Demand for data skills is strong and growing: data work is one of the fastest-growing skill areas around. But the role is being rebuilt from the bottom by AI: tools now write SQL from plain English, build dashboards, clean data, and generate reports, automating exactly the entry-level tasks the job was built on. So the work is shifting up, from running queries to framing the question and judging whether the answer holds.
Right now
It’s a growing field, but the bottom rung is eroding: the routine SQL, dashboard, and reporting work that used to be how juniors broke in is the most-automated layer, so a pure "I can run queries" profile is weakening even as overall data demand rises. The field isn’t shrinking, but the easy way in is narrowing, and the analysts who do well use the tools to skip the busywork and get fast to the judgment.
Sources: BLS OEWS (Operations Research Analysts; Data Scientists +34% 2024-34, May 2024); WEF 2026 Future of Jobs (AI + big data top skill area); BI-copilot / text-to-SQL coverage (2025-26). Dated June 2026.
Would you actually like it?
Worth a look if chasing the why behind a number sounds more fun than tedious.
In practice, people realize it’s their thing when…
- they like turning a fuzzy question into a sharp one that the numbers can actually answer
- they’re skeptical by habit, they want to check whether a number is really telling the truth
- they do not mind the unglamorous tidying up and digging that gets to the answer
- they like being near the real decisions, turning a pile of numbers into something someone will act on
…and it probably isn’t their thing when
- they pictured a run of clever discoveries, and most of it is tidying up messy numbers and building the same charts again every week
- they want an easy way in: pulling numbers out and putting them on a screen used to be the beginner job, and computers do that now
- they only want to get numbers out and turn them into charts: that exact part is what computers now do, so what you have to bring is the thinking
Think Like a Data Analyst: Find a Story in Real Data
Pick a question you actually care about, find real numbers anyone can download that could answer it, dig in, and write something people can read that argues what you found, with a couple of charts. The hard part, and the whole point, is asking the right question and checking the answer is real, not running the code.
Or try one of these