What data analysts do
Understand what a data analyst does day to day, how the work flows from question to insight, how it differs from related roles, and which skills to build first.
A data analyst helps people make better decisions using data. A shop owner asks, "Why did sales drop in March?" A product manager asks, "Which feature do users ignore?" A bank asks, "Which branches are slow at approving loans?" The analyst finds the data, cleans it, answers the question and explains the answer in a way a busy person can act on.
The job is mostly about questions and communication, not about fancy mathematics. If you can use a spreadsheet, write simple SQL and explain a chart clearly, you can start.
What the work looks like
Think of an analyst as a translator between a business question and the data that can answer it. A typical project has five steps.
The analyst workflow
- 1Ask: turn a vague request into a precise question. "Sales are bad" becomes "Did weekday evening revenue fall compared with the previous four weeks, and in which items?"
- 2Get the data: find where it lives (a database, an export, a Google Sheet, an app) and who owns it.
- 3Clean it: fix duplicates, missing values, wrong types and inconsistent spellings. This often takes the largest share of the time.
- 4Analyse: compute totals, averages, trends and comparisons with spreadsheets, SQL or Python.
- 5Communicate: present a chart, a dashboard or a short written note that says what happened, why it matters and what to do next.
A day in the life: Brew & Bean
Suppose you are the analyst for Brew & Bean, a coffee shop in Indiranagar, Bengaluru. The owner says: "Weekday evenings feel quiet. Should I close at 9 pm instead of 10 pm?"
You would pull the orders table, count orders and revenue by hour of day, compare the 9 pm to 10 pm hour with the cost of staying open, and check whether those customers return at other times. Your answer might be: "The 9 to 10 pm hour brings about X orders on weekdays and mostly takes customers who would otherwise come at 8 pm. Closing at 9:30 pm is worth testing for a month." Note that the analyst recommends and the owner decides.
Always start with the decision
Before opening any data, ask: "What decision will this answer change?" If nobody can name a decision, the analysis is probably not worth doing.
Types of analytics
Four kinds of questions
| Type | Question | Example at Brew & Bean |
|---|---|---|
| Descriptive | What happened? | Revenue last week was lower than the week before |
| Diagnostic | Why did it happen? | Cold Brew sales fell because it was out of stock on two days |
| Predictive | What is likely to happen? | Expected orders next Saturday, based on previous Saturdays |
| Prescriptive | What should we do? | Stock more cold brew ingredients on Fridays |
Most entry-level analyst work is descriptive and diagnostic. Predictive work leans towards data science and machine learning; see the Data Science and ML track if that interests you.
Where analysts work and what they are called
Almost every organisation with customers or operations has analysts: IT services and product companies, e-commerce, banks, insurance, healthcare, logistics, telecom, retail, government and start-ups. Titles overlap a lot, and a job post's title matters less than its tasks.
Related roles
| Role | Main focus | Typical tools |
|---|---|---|
| Data analyst | Answers questions from data, builds reports and dashboards | Excel, SQL, Power BI or Tableau, sometimes Python |
| Business analyst | Understands business processes and requirements; uses data to support them | Excel, SQL, documentation and process tools |
| Product analyst | Studies how users behave in an app or website | SQL, product analytics tools, A/B tests |
| Data engineer | Builds the pipelines that move and store data reliably | SQL, Python, cloud platforms |
| Data scientist | Builds statistical and machine learning models | Python, statistics, ML libraries |
Skills you need
Skills and where this track teaches them
| Skill | Why it matters | Where to learn it |
|---|---|---|
| Spreadsheets | Fast exploration and small reports | Spreadsheets for analysis |
| SQL | Pulling data from databases | SQL for analysts |
| Data cleaning | Real data is messy | Cleaning messy data |
| Metrics and KPIs | Knowing what to measure | Business metrics and KPIs |
| Dashboards | Sharing numbers that update automatically | Dashboards with Power BI and Looker Studio |
| Communication | People act only on what they understand | Data storytelling and presenting insights |
| Basic statistics | Avoiding wrong conclusions | A/B testing and experiments |
Python (pandas) is useful and common, but many entry-level roles start with Excel and SQL. Learn those two well first. A domain background also counts: a commerce graduate who understands accounting, or a B.Pharm graduate who understands pharma sales, is valuable.
Common misconceptions
Myths that put beginners off
"You need a maths degree." You need arithmetic, percentages, averages and logical thinking. Statistics comes with practice.
"Analysts only make charts." Charts are the last 10 percent. The value is in asking the right question and checking the data.
"AI will replace analysts." AI tools help write queries and summarise tables, but someone must know the business, check the numbers and take responsibility for the answer. See *Using AI as an analyst, and checking its work*.
How this comes up in interviews
Interviewers test three things: can you handle data (SQL and Excel questions), can you think (a business case such as "orders dropped 10 percent, how would you investigate?"), and can you explain (walk through a past project). For the case question, a safe structure is: clarify the metric and the time period, check the data for errors, split the drop by segment (product, location, channel, new or returning customers), form a hypothesis, then say what you would check next.
If you have no work experience, build two or three projects with public data and explain them clearly. The last lesson of this track is a project you can adapt.
Try it: Rewrite vague requests as analyst questions
Rewrite each request as a precise question that includes a metric, a time period and a comparison. Then name the data you would need. 1. "Customers seem unhappy." 2. "Is our new menu working?" 3. "Which marketing is best?"
Act as an interviewer for a [junior data analyst] role at a [company type, e.g. food delivery start-up]. Ask me one business case question at a time, such as "Orders fell 10 percent last week. How would you investigate?" After each of my answers, tell me what I did well, what I missed, and ask a follow-up. Keep your feedback under 100 words. Start now.What to expect: A good session makes you clarify definitions, split the metric by segment and suggest concrete checks. Treat the feedback as practice, not as an authority on what a particular employer wants.
Key takeaways
- An analyst turns business questions into data answers people can act on.
- The workflow is ask, get data, clean, analyse, communicate; cleaning often takes the most time.
- Descriptive and diagnostic questions make up most entry-level work.
- Start with Excel and SQL; add dashboards, statistics and later Python.
- Analysts recommend; the business decides. Always start from the decision.
Check yourself
Finishing the quiz marks the lesson as done.
1. Which is the best first step when a manager says "sales feel low"?
2. "Cold Brew sales fell because it was out of stock on two days." Which kind of analytics is this?
3. Which role mainly builds and maintains the pipelines that move data into databases?
4. Which tools are the most common starting point for entry-level analyst roles?
5. In an interview case "orders dropped 10 percent", what is a strong early move?
Saved on this device. Create a free account to keep it on every device.
Practise this with Sam
A real mock interview, out loud, with an honest scorecard. Free to start.