Data and analytics · 5 stages · about 42 weeks
Data analyst roadmap
0/25 tasks done (ticks are kept on this device)
- 1
Beginner · about 6 weeks
Excel and spreadsheets
What to learn
Spreadsheet basics: Excel is the most used data tool in Indian offices.
Learn rows, columns, cells, sheets and how to enter and format data. Keep one table per sheet, with one header row and one type of value per column. Clean, simple tables make every later step easier.
Core formulas: Formulas do the maths for you and update by themselves.
Start with SUM, AVERAGE, COUNT, MIN, MAX and ROUND. Then learn IF, COUNTIF, SUMIF and their multi-condition forms COUNTIFS and SUMIFS. Learn absolute references with $ so a formula copies correctly.
Lookup formulas: Real data is spread across several tables.
XLOOKUP (or VLOOKUP in older Excel) finds a value in another table, like the price of a product code. INDEX with MATCH does the same in a more flexible way. Learn to handle values that are not found, so you do not get #N/A everywhere.
Text and date functions: Raw data often has messy names and dates.
TRIM removes extra spaces, PROPER fixes capital letters, and LEFT, RIGHT and TEXTSPLIT pull out parts of text. DATE, MONTH, YEAR and EOMONTH help you group sales by month. Watch out for dates typed in different formats, like 05/04 meaning April or May.
Sorting, filtering and conditional formatting: These help you see what matters quickly.
Sort to find the top or bottom values, and filter to show only some rows, like one city. Conditional formatting colours cells, for example red for stock below 10. Turn your data into an Excel Table so filters and formulas grow with the data.
Pivot tables: Pivot tables summarise thousands of rows in seconds.
A pivot table groups data and totals it, for example sales by month and by product. You drag fields into rows, columns, values and filters. This is one of the most asked-for skills in data analyst interviews.
Charts: A good chart shows a pattern faster than a table.
Use a line chart for change over time, a bar chart to compare categories, and avoid 3D and crowded pie charts. Give every chart a clear title and labelled axes. Ask yourself what one message the chart should show.
Tasks
Project
A one-page Excel dashboard for a small shop: total sales, sales by month, top 5 products, GST collected, and low-stock alerts, built with formulas, a pivot table and 3 clear charts.
Free resources
- 2
Beginner · about 8 weeks
Statistics basics and SQL
What to learn
Descriptive statistics: Statistics tell you what is normal and what is unusual in data.
Mean is the average, median is the middle value and mode is the most common value. Standard deviation shows how spread out the data is. When a few very large values exist, like salaries, the median is often more honest than the mean.
Distributions and percentiles: They show the shape of your data.
A histogram shows how many values fall in each range. Percentiles tell you, for example, that 90 percent of orders are below ₹2,000. Learn what a normal (bell-shaped) distribution is and what skewed data looks like.
Correlation is not causation: This one idea saves you from many wrong conclusions.
Correlation means two things move together, like ice cream sales and AC sales in summer. It does not prove one causes the other; both may be caused by heat. Learn the basics of A/B tests, which are the fair way to test a cause.
SQL SELECT and filtering: Most company data lives in databases, and SQL is how you ask for it.
SELECT picks columns, FROM picks the table, WHERE filters rows and ORDER BY sorts them. Learn LIMIT, DISTINCT and simple conditions with AND, OR, IN and LIKE. SQL is asked in almost every data analyst interview.
GROUP BY and aggregates: Business questions are usually totals and averages.
COUNT, SUM, AVG, MIN and MAX work with GROUP BY to give totals per group, like sales per city. HAVING filters those groups. This is the SQL version of a pivot table.
JOINs: Data is split across many tables.
A JOIN combines tables using a shared column, like customer_id. INNER JOIN keeps only matching rows, and LEFT JOIN keeps every row from the first table. Draw the tables on paper first and check your row counts after each join.
Subqueries, CTEs and window functions: These answer harder questions in one clear query.
A CTE (WITH ... AS) names a step of your query so it is easier to read. Window functions like ROW_NUMBER, RANK and running SUM let you rank products or show month-on-month growth. These often separate strong candidates in interviews.
Tasks
Project
A SQL case study on a public e-commerce or retail dataset: 15 business questions, the SQL for each, the answers, and a one-page summary of what the business should do next, shared on GitHub.
Free resources
- 3
Intermediate · about 8 weeks
Power BI or Tableau: dashboards and data storytelling
What to learn
Power BI basics: Power BI is one of the most asked-for tools in Indian analyst jobs.
Power BI Desktop is free on Windows. You load data, clean it, build a data model and make interactive reports. Reports can then be shared through the Power BI service at work.
Power Query: Most time in analysis goes into cleaning data.
Power Query (in both Excel and Power BI) cleans data in repeatable steps: remove columns, fix types, split text, merge tables and unpivot. When new data arrives, you just press refresh. This saves hours of manual copy and paste every month.
Data modelling: A good model makes reports correct and fast.
Use a star schema: one fact table with events like sales, linked to dimension tables like products, customers and dates. Always add a proper date table. Check relationships carefully, because wrong ones give wrong totals.
DAX measures: DAX lets you calculate business numbers like growth and margin.
DAX is the formula language of Power BI. Start with measures like Total Sales = SUM(Sales[Amount]), then learn CALCULATE and time functions for month-on-month and year-on-year growth. Prefer measures to calculated columns for totals.
Tableau basics: Some companies, especially larger ones, use Tableau instead.
Tableau builds charts by dragging dimensions and measures onto rows and columns. Tableau Public is free and lets you publish dashboards online for your portfolio. The ideas are the same as Power BI, so learning one makes the other easier.
Dashboard design: A dashboard is useless if people cannot read it in a minute.
Put the most important numbers (KPIs) at the top, then trends, then details. Use few colours and the same colour for the same thing everywhere. Add filters people really need, like month, city and product.
Data storytelling: Your job is to help people decide, not just to show charts.
Start with the question, show the key finding, then give the reason and a clear suggestion. Remove anything that does not support the message. Practise explaining one chart in 30 seconds in simple words.
Tasks
Project
An interactive sales dashboard for a retail chain (sales, profit and GST by month, city and category, with top and bottom products) in Power BI or Tableau, plus a one-page summary of 3 findings and 3 suggestions.
Free resources
- 4
Intermediate · about 10 weeks
Python with pandas and data cleaning
What to learn
Python basics: Python handles data that is too big or too messy for Excel.
Learn variables, lists, dictionaries, loops, if/else and functions. Use Jupyter Notebook or Google Colab, where you run code in small cells and see the result at once. You do not need to be a software developer, but you need to be comfortable.
pandas DataFrames: pandas is the main Python tool for tables of data.
A DataFrame is like an Excel sheet inside Python. Learn to read CSV and Excel files, select columns, filter rows and sort. head(), info() and describe() give you a quick first look at any dataset.
Data cleaning: Real data is always messy, and wrong data gives wrong answers.
Find and handle missing values, remove duplicates, fix data types and clean text like city names typed in different ways (Bengaluru and Bangalore). Check for impossible values, like negative ages. Write down every cleaning step so others can trust your work.
Grouping, merging and reshaping: These are the pandas versions of pivot tables and lookups.
groupby gives totals per group, merge joins tables like SQL, and pivot_table reshapes data like Excel. Learn to work with dates using to_datetime and resample by month. Check the row count after every merge.
Visualisation with Python: Charts help you explore and explain data.
Matplotlib and seaborn draw line, bar, histogram and scatter charts from DataFrames. Use charts first to explore (find patterns and problems), then to explain. Keep the same rules as in Excel: clear titles, labels and one message per chart.
Exploratory data analysis (EDA): EDA is how you understand a new dataset before answering questions.
Look at the size, columns, missing values, ranges and distributions. Then ask simple questions and answer them with groupby and charts. Write short notes on what you find as you go.
Reproducible notebooks: Good analysis can be rerun and checked by others.
Order your notebook from top to bottom: load, clean, analyse, conclude. Add short text cells that explain each step. Restart and run all cells before sharing, to be sure it works.
Tasks
Project
An exploratory analysis of a real Indian public dataset (for example rainfall, crop prices or a city transport dataset) in a Jupyter notebook: cleaning, 6 to 8 charts, clear findings, and a short README on GitHub.
📘 Learn it here, free: Python Programming course
Free resources
- 5
Expert · about 10 weeks
Business understanding, advanced analysis and your portfolio
What to learn
Business metrics and KPIs: Analysts who understand the business are much more useful.
Learn common metrics: revenue, profit margin, customer acquisition cost, retention, churn, average order value and conversion rate. Know which ones matter for the business you work with, like a shop, a bank or an app. Always ask what decision a number will help make.
Asking the right question: A clear question saves days of wasted analysis.
Before starting, write the business question, who will use the answer and what they will do with it. Break big questions into small ones you can measure. Agree on definitions, like what counts as an active customer.
Cohort, funnel and retention analysis: These show how customers behave over time.
A funnel shows how many users move from one step to the next, like visit, add to cart and pay. A cohort groups customers by the month they joined and tracks how many come back. These are common in product and e-commerce companies.
A/B testing and basic inference: Companies test changes before rolling them out.
An A/B test shows version A to some users and version B to others, then compares results. Learn sample size, p-values and confidence intervals at a simple level. Do not stop a test early just because the numbers look good.
Automation: Repeating the same report by hand wastes time and causes mistakes.
Automate monthly reports with Power Query refresh, scheduled Power BI refresh or Python scripts. Keep a checklist to check numbers after each run. Spend saved time on deeper analysis.
Data ethics and privacy: You will often handle personal data.
Use only the data you need and hide personal details like phone numbers when you can. Learn the basics of India's Digital Personal Data Protection Act and your company's rules. Never share raw customer data outside approved tools.
Portfolio and job search: A portfolio proves your skills better than a certificate.
Build 3 to 5 strong projects that answer real business questions, each with a short README, the data source, your method and your findings. Share them on GitHub, Kaggle or Tableau Public and link them on your CV and LinkedIn. Practise SQL and Excel interview questions often.
Tasks
Project
A complete case study for a business: define the problem, get and clean the data with SQL and Python, build a Power BI or Tableau dashboard, and write a 2-page report with findings and recommendations, all published in your portfolio.
Free resources