What does a data analyst actually do?
Companies hold a huge amount of data about sales, customers and their websites, but that data has to be understood before it can guide decisions. A data analyst collects this data, cleans it, looks for patterns in it and presents the findings to the team as charts or dashboards, so the company can make the right call.
The good news is that you do not need to be an expert in everything to get started. By learning a few basic tools in the right order, you can begin working at an entry level.
Data Analyst vs Data Scientist
The two are often mistaken for each other, but the work is different.
| Data Analyst | Data Scientist | |
|---|---|---|
| Main work | Pulling insights out of existing data and building reports | Building machine learning models to predict the future |
| Tools | Excel, SQL, Power BI, Tableau | Python, ML libraries, advanced statistics |
| Maths | Basic statistics is enough | Statistics and linear algebra in depth |
| Getting started | An easier path into a first career | Needs more study and practice |
How do the three roles differ?
Many people treat these as the same job, but the work, the skills and the amount of interaction with people are different in each.
| Business Analyst | Data Analyst | Data Scientist | |
|---|---|---|---|
| Main work | Understanding the business problem and giving recommendations from data findings | Cleaning and analysing data and drawing out insights | Building ML and mathematical models |
| Coding | Little to none | Moderate: SQL and basic Python | Deep: Python, ML |
| Main tools | Excel, Power BI, PPT | Excel, SQL, Power BI, Python | Python, statistics, ML libraries |
| Interaction with people | A lot: clients, managers, stakeholders | With the team, to some extent | Less, mostly with the technical team |
In simple terms: the Business Analyst is the front-facing (client side) role, the Data Scientist is the behind-the-scenes (technical side) role, and the Data Analyst is the bridge between the two. A Data Analyst is usually expected to know everything a Business Analyst does, plus some coding.
Is a degree necessary?
A degree, especially from a good college, helps at the start, and some companies even filter by CGPA or college. But once you have a job, it is the quality of your work and your skills that move you forward.
If you do not have a B.Tech, BCA or MCA, you can still enter this field on the strength of good projects and skills. The path may demand a bit more effort, but it is not closed. And if you do want a degree too, online programs such as the IIT Madras BS in Data Science are available without JEE.
Step-by-Step Roadmap
-
Start with Excel
This is the most basic tool of every data analyst. Learn formulas, filters, sorting, removing duplicates, Text to Columns and Pivot Tables. Pivot Tables feel hard at first, but they save the most time. Building charts and remembering keyboard shortcuts also speeds up your work.
-
Learn SQL
Most companies keep their data in databases, and SQL is the language used to pull data out of them. You should be comfortable with SELECT, WHERE, GROUP BY, JOIN and subqueries. MySQL is an easy option to begin with, and moving to PostgreSQL or SQL Server later is not difficult.
Try our free SQL course in English -
Basic understanding of statistics
Mean, median, mode, average, probability and conditional probability: this much you must know. You do not need PhD-level maths, but you should understand what your numbers mean. The ability to read graphs properly comes in handy here too.
-
Power BI (then Tableau)
Power BI is a good place to start for turning data into dashboards, because it connects easily with other Microsoft products (Excel, Azure). After that you can learn Tableau as well. Which tool suits which project is something you understand with experience.
-
Python basics and Pandas
Some tasks become very long or difficult in Excel. A little Python helps there. With the Pandas library you can read Excel or CSV files, clean data and save the result back to Excel. In an analyst role you do not need advanced programming, just enough to get your work done.
-
Build projects and prepare your resume
Learning tools alone is not enough. Build 3 to 4 real projects, such as a sales analysis of a public dataset or a dashboard, and keep them on GitHub or in a portfolio. In your resume, instead of listing tools, write what result you got from using them.
Use our Resume Builder
How AI has changed learning data analytics
Earlier, getting stuck on a new tool or programming language meant hours of debugging. Today, AI assistants like ChatGPT and AI-powered spreadsheet tools can write your formula, explain an error and prepare a first draft of Python or SQL. This speeds up learning a great deal, especially for people who are new to coding.
But AI is a helper, not a replacement. See below where it helps and where you still have to think for yourself.
| What AI does well | What you have to do yourself |
|---|---|
| Explaining Excel formulas and Pivot Table steps | Choosing the right question about what to learn from the data |
| Drafting SQL queries and Pandas code | Checking whether the output is correct |
| Explaining error messages | Understanding data quality and business context |
| Suggesting chart ideas and a report outline | Making decisions based on the findings |
Why basics still come first
AI sometimes gives a wrong formula, or code that runs but produces the wrong result. To catch this, you need a basic understanding of Excel, SQL and statistics. That is why in the plan below, each month starts with the basics and only then brings in AI's help.
A 4-Month Plan
This is a suggested sequence. You can stretch or shorten it according to your pace and available time.
-
Month 1: Excel and Data Cleaning
Learn filters, sorting, Text to Columns, removing duplicates, formulas and Pivot Tables. Many people think they already know Excel, but these basics are often missed. When a formula does not come to mind, ask AI, but try on your own first.
-
Month 2: SQL and basic Python
Get SELECT, WHERE, GROUP BY and JOIN solid in SQL. In Python, learn the basics of variables, loops and functions, then learn to read Excel or CSV files with Pandas. Here, let AI write code, but read every line and understand what it does.
Free SQL course in English -
Month 3: Statistics and Dashboards
Learn the basics of mean, median, mode and probability. Then learn to turn data into a dashboard with Power BI, and after that, if you like, look at Tableau too.
-
Month 4: Advanced Excel, Git and Projects
Learn Advanced Excel (such as Power Query and macros) and the basics of Git and GitHub, so you can keep your work safe and presentable. Then build 3 to 4 real projects and put them on GitHub. Once a project is done, write its summary in your resume.
Use the Resume Builder
Tips for getting better results from AI
- Give full context. Tell it the column names, a small sample of the data and the result you want.
- Ask for one thing at a time. Breaking a big question into small steps gives more accurate answers.
- Check the result yourself. Work it out by hand on a small sample and see whether AI's answer matches.
- Do not paste sensitive data. Never paste a company's or its customers' private data into AI tools.
- Learn with understanding. Instead of just copying code, understand it, or you will get stuck in the interview.
How to build a good project
Better than a random project is one that tells a complete story from start to finish. Free datasets are easy to find on Kaggle.
Pick a problem
For example, "Which month and which product have the highest sales?"
Clean the data
Remove missing values and duplicates, using SQL or Pandas.
Analysis and Dashboard
Find patterns and show them in Power BI or Tableau.
Write the insights
Not just a chart: explain the story behind it. A business wants recommendations, not numbers.
Put it on GitHub
Write a README covering the problem, the approach and the result.
Who is your resume written for?
Your resume is read in two ways. First, software called an ATS (Applicant Tracking System) often scans it and matches it against the job's keywords. After that, a recruiter looks at it for a very short time. So your resume should make sense to both the machine and the human.
Keep in mind that a good resume improves your chances of getting an interview, but it does not guarantee a job.
4 things that get a resume shortlisted
-
Role-specific headline
A line like "Fresher looking for opportunity" will not get you found in a recruiter's search. Instead, write the role and tools, such as Data Analyst | SQL | Excel | Power BI | Python. Recruiters search with exactly these words.
-
Projects: your proof
No degree proves that you can do the work; your work does. For every project, write four things: what the problem was, what the data looked like (how many rows), what analysis you did, and what insight came out of it. See three examples below.
Weak Better Used Power BI on sales data. Cleaned retail sales data of 50,000 rows with SQL (removed about 3% duplicate and empty rows) and built a Power BI dashboard showing trends by region and product. Result: just 20% of products were bringing in about 70% of revenue. Did an analysis of customer data. Identified repeat customers in 12,000 e-commerce orders using Python (Pandas). Result: only 18% of customers bought again, but they accounted for about 40% of total revenue. Suggestion: a loyalty offer for these customers. Looked at marketing data in Excel. Built Pivot Tables and charts on 5 months of campaign data (8,000 leads). Result: Campaign B cost ₹180 to bring in one customer, versus ₹320 for Campaign A. Suggestion: shift budget toward B. The formula is simple: what you did + how much data + what result you got + what you recommended. The numbers above are only examples. In your resume, write only the real numbers from your own project (how many rows, what percentage, what difference), and never make up a number you did not actually work out, because that is exactly what the interview will ask about.
-
ATS keywords
Read the job description carefully. Use in your resume the same words for the tools and skills it lists, as long as you truly know them (SQL, Excel, Power BI, Python and so on). Do not write what you do not know; you will get caught in the interview. Keep the format simple: a single column, clear headings (Skills, Projects, Education), and avoid tables or images, because ATS cannot read them properly.
-
Clickable links
Put your GitHub, portfolio and LinkedIn profile in your resume, so the recruiter can see for themselves that you really did the work. Keep your LinkedIn headline the same as the headline on your resume.
What is domain specialization and why does it help?
A domain is the field whose data you work on, such as banking, healthcare, marketing or e-commerce. The tools stay almost the same everywhere (SQL, Excel, Power BI), but what the data means is different in each place.
If you have a basic understanding of that field, you can ask the right questions and explain the findings in business language. When a company looks for an analyst for a particular domain, a resume with projects and understanding from that same domain stands out.
Four common domains at a glance
| Domain | Typical data | Typical questions |
|---|---|---|
| Finance and Banking | Transactions, loans, customer accounts | Which customers are at risk of struggling to repay a loan? What are the patterns of fraud? |
| Healthcare | Hospital operations, appointments, patient records | How can waiting time be reduced? How well are beds being used? (Data privacy is extremely important here) |
| Marketing | Campaigns, website and ads data | Which campaign brought in more customers? How should customers be split into groups? |
| E-commerce and Retail | Orders, inventory, returns | Which products sell the most? Which customers buy again? |
How to choose your domain
-
Look at your interests and background
It is easier to start with a subject you enjoy or already have a little experience in. For example, someone with an ECE background may take an interest in telecom or electronics manufacturing data.
-
Look at job postings
On LinkedIn or Naukri, search "Data Analyst" along with different domain names and see where there are more openings and what skills they ask for.
-
Stick to one domain
Choose one domain at the start. Moving to another later is always possible, but showing a little of everything at the beginning does not make anything stand out.
What to do after choosing
Learn the basic vocabulary of the domain
Read the common terms and metrics of that field, for example in company reports, blogs and industry news.
Build 1 to 2 domain-specific projects
Find free datasets for your domain on platforms like Kaggle and build a project that tells a complete story.
Show the domain in your resume and LinkedIn
Put the domain name in your headline and projects, and write results that fit it.
Look for an internship or small gigs
Freelance projects or internships give real experience, which carries the most weight on a resume.
Why don't you get a job even after learning the skills?
In most cases the shortfall is not in skills. Three things often become the obstacle.
-
Your resume does not inspire trust
A recruiter has only one page of yours. You have to convince them that you can do this work. Ways to do that: good internships, real projects, GitHub, and a credible course or community you learned from. Even if your college is not a big name, these things can build your credibility. If you are from a Tier 3 college, also see our placement guide.
-
Many tutorials, few projects
You do not learn a tool by watching hundreds of videos; you learn it by working with it. The one who has never practised is the one who gets stuck in an interview or task. Build at least one project on every tool, and write what problem you solved and what result came out of it. Use the Resume Builder.
-
Not knowing how to ask for a referral
Nobody replies to a message like "Bro, get me a job." Make the other person's work easy: give the job ID, attach your resume, and if you and they have anything in common, mention it in one line. A simple template is below.
Referral message template
Hi [Name], I'm [Your Name], a [B.Tech/BCA/...] graduate with skills in SQL, Excel and Power BI. I saw the Data Analyst opening at [Company] (Job ID: [XXXX]). I recently built [one-line project, e.g. a sales dashboard in Power BI]. My resume is attached. If you feel I'm a good fit, would you be open to referring me? Any advice would also be appreciated. Thank you for your time.
Keep the message short, tailor it for each person, and do not take it personally if there is no reply.
What is asked in interviews
Interviews usually test two things: technical understanding and the ability to explain your point.
- SQL scenario questions. For example: find the total revenue of each customer and show the top 3. How would you calculate month-over-month sales growth? How would you find duplicate records? Practise GROUP BY, JOIN and window functions for these. Free SQL course
- Explaining your project. Say in one minute what the problem was, what you did and what the result was. Instead of listing tool names, emphasise the outcome.
- Clear communication. Explain technical things as if you were telling someone who does not know data. Practise by doing a mock interview with a friend.
A straight word: the path is not easy for freshers
For a first job you often have to send a great many applications. Some get shortlisted, some lead to interviews, and then somewhere it works out. Stopping after 10 to 15 applications is not right. Keep learning, building projects and connecting with new people.
The good news is that data and business analysts are needed in almost every kind of company, and people can come into this field from sales or any other field too.
Frequently asked questions
How long does it take to become a data analyst?
It depends on your practice. With regular daily effort you can be ready for an entry-level role in a few months, but the real polish comes from projects and experience.
Can you become a data analyst without coding?
Yes, you can start with Excel, SQL and Power BI. Learning Python later widens your options.
Will AI make data analyst jobs disappear?
AI is becoming part of Excel and BI tools, which means it speeds up the work. The analyst who learns to use AI in their workflow stays ahead.