Understand what a Data Analyst does and the value they create.
Goal: Understand what a Data Analyst does and the value they create.
Nadia's first week at Perch, the online furniture retailer, ends with a message from Marcus, the Head of Marketing: "Quick one — how are sign-ups doing?"
She stares at it. Sign-ups for what? Doing well compared to what? Last week, last year, the goal nobody wrote down? Marcus doesn't want a spreadsheet. He's about to decide where next month's ad budget goes, and he wants Nadia to tell him whether sign-ups are healthy enough to keep spending the same way.
That gap — between a pile of numbers and a person who has to decide something — is the entire job. A Data Analyst's one core job is to turn raw data into answers that people act on. Perch records everything: every sofa sold, every desk page viewed, every email opened, every checkout abandoned. All of it sits in tables, meaning nothing, until someone shapes it into "here's what's happening, and here's what I'd do about it."
Nadia ran weekly sales reports for years at her old retail job. She already knows the feeling of numbers piling up faster than anyone can read them. What's new is that her whole role now is the bridge from those numbers to a choice someone makes.
Raw data is what the company collected. The answer is what someone can finally do with it. The analyst is the bridge.
Hold onto that bridge. Everything else in this course — spreadsheets, SQL, charts — is just how you build it stronger.
Here's a line Nadia gets wrong in her first month, and it's worth getting right early.
She pulls the sign-up numbers, sees they're down 12% from the month before, and almost writes back: "We should cut the Facebook ads." Then she stops. That's not her call.
Picture a ship at sea. The data is everything the instruments record — speed, depth, fuel, the weather closing in. The analyst is the navigator who reads those dials and says, "We're drifting off course, and here's the fastest way back." The captain — at Perch, that's Marcus or Dana, the VP of Operations — is the one who turns the wheel. The navigator who's never wrong about the instruments is still not the person steering the ship.
So Nadia rewrites her reply. Sign-ups are down 12%, and the drop is almost entirely from one ad campaign that changed its landing page on the 3rd. Here's the evidence; here's what I'd look at first. Marcus makes the actual budget decision. He decides better because she handed him the map.
This is freeing, once it clicks. Nadia isn't on the hook for Perch's revenue. She's on the hook for the answer being true and clear — for Marcus understanding exactly what the data shows before he bets money on it. The analyst supplies the evidence; the manager makes the call.
Notice what Marcus actually asked. Not "calculate the standard deviation of weekly registrations." He asked "how are sign-ups doing?" — a business question wearing everyday clothes.
This is the part that surprises career-changers. Nadia expected the job to be mostly math. It's mostly translating fuzzy business questions into things data can answer. At Perch, trying to grow repeat customers, the real questions sound like:
Each one is a decision waiting to happen. If shelving drives repeat purchases and sofas don't, Perch knows what to feature on the homepage. If 40% of carts die on the shipping-cost screen, someone has a reason to rethink shipping.
The statistics and the SQL are in service of these questions — never the other way around. Priya, Nadia's mentor, has one rule she repeats until it sticks: define the question before you pull a single row. An analyst who can run every calculation but can't say which business decision it informs has done arithmetic, not analysis.
So what does Nadia actually do between 9 and 5? A normal Tuesday at Perch looks like this.
Morning. A request lands, usually vague. Marcus again: "Are the new customers from the spring campaign any good?" Before opening any tool, Nadia messages back to pin down what "good" means: do they come back and buy again, or just buy once and vanish? She and Marcus agree that a "good" customer places a second order within 60 days.
Mid-morning. Now she pulls the data. Sometimes that's exporting a sheet from Perch's email platform; today it's a SQL query against the company's data warehouse to get every spring-campaign customer and their order history. When a table is missing a column she expects, she pings Tom, the data engineer who owns the pipelines, instead of guessing.
Midday. The raw data is messy — duplicate rows, a few customers with blank signup dates, prices stored as text in one column. She cleans it, then explores: counts, averages, a quick look at who reordered and who didn't. Most of the time goes here, and almost none of it is glamorous.
Afternoon. She builds a chart — repeat-purchase rate for spring customers versus everyone else — in a BI tool so Marcus can see it without reading a table. Then the most important step: she writes the "so what." Spring customers reorder at 18%, below our 25% baseline; they came cheap but they don't stick. I'd shift budget toward the channels that bring repeat buyers.
Question, data, cleaning, chart, "so what." That last sentence is what Marcus remembers. Everything before it was plumbing.
Step back from Nadia's Tuesday and you can see the shape of every analysis she'll ever do. It's a repeatable loop:
question → data → cleaning & analysis → insight → decision → (often) a new question.
That last arrow matters. Good answers breed sharper questions. Nadia's finding — spring customers don't stick — sends Marcus straight back with a better one: "Then which channels DO bring repeat buyers?" The loop turns again. This is why analysts are never "done"; they're a flywheel the business keeps spinning.
Now, the tools. A junior analyst today touches three things almost daily:
Python is a strong plus, not a requirement to start. And here's the thing to internalize before any of it: the tool is the route, not the destination. Nobody at Perch will ever praise Nadia for "a really elegant SQL query." They'll thank her for telling them, in one clear sentence, what to do next.
That's also why the skill travels. Marketing, product, finance, operations — every team at Perch and at almost every company needs evidence-based answers, and the loop is the same wherever you point it. Learn it once on furniture sales and you can run it on hospital admissions or ad clicks. The domain changes; the job doesn't.
Back to Marcus's very first message. Watch the loop carry a vague question all the way to a decision.
Question. "How are sign-ups doing?" Nadia doesn't pull anything yet. She asks two clarifying questions and lands on a real one: Are weekly new-account sign-ups up or down versus the last 8 weeks, and is any single channel driving the change? Now it's answerable.
Data. She writes a SQL query against Perch's warehouse for weekly sign-up counts, split by where each person came from (search, email, paid ads). The "paid ads" numbers look impossibly low for two weeks, so she checks with Tom — a tracking tag had broken at the source. Without that catch, her whole answer would have been wrong.
Cleaning & analysis. She drops the corrupted rows, confirms the rest ties out to the email platform's own totals, and compares each week. Sign-ups are down 12% over two weeks, and nearly all of the drop is in one paid campaign.
Insight. She builds a single line chart and writes the "so what": Overall sign-ups fell 12%, but it's not broad — one paid campaign collapsed after its landing page changed on the 3rd. Everything else is flat to up.
Decision. Marcus doesn't cut the budget across the board, which is what he'd have done from the scary headline number. He pauses the one broken campaign and asks design to revert the landing page. Dana, the VP, gets it in one sentence in the Monday review: "Sign-ups dip is one fixable campaign, not a trend."
New question. A week later: "Did reverting the page bring those sign-ups back?" The loop turns again.
One fuzzy request became a clear, true answer that changed what the company did — and that is the whole job.
Think of any decision someone near you made recently — at work, at home, anywhere ("should we switch suppliers?", "which product should we discount?"). Write it as a single business question. Then list, in plain words, the data you'd need to answer it and the one-sentence "so what" you'd hope to hand the decider. You've just walked the analyst loop without touching a single tool — which is the point of this whole topic.
Preparing your quiz…