One field, several very different jobs
Job boards use "data science" for everything from building Power BI dashboards to deploying recommendation models. The skills overlap at the start and then separate sharply, which is why a single list of topics confuses most learners.
This roadmap shows the stage everyone shares, the branch each role takes, and where each branch leads. For a month-by-month plan to your first application, read our guide on how to become a data scientist in India.
Stage 2: the branch for each data role
Each branch matches a role page. The foundation stays useful in all of them.
Data Scientist
Supervised and unsupervised learning in Scikit-learn, feature engineering, cross-validation, and explaining results with precision, recall and business impact. First proof: a modelling project with a clear write-up of what the model means.
ML Engineer
Take models to production: Flask or FastAPI, Docker, Streamlit demos and basic AWS or Azure deployment, plus monitoring. First proof: a model others can call over an API.
Data Engineer
Go deeper on SQL, data modelling and ETL pipelines that load, clean and schedule data reliably. First proof: a pipeline that refreshes a database from raw files every day.
BI Analyst
Power BI and Tableau dashboards, DAX basics and KPI definitions agreed with business users. First proof: a dashboard a manager could use without your help.
Data Analyst
Strong SQL, Excel and visual storytelling for recurring business questions. First proof: an analysis that ends in a clear recommendation.
Analytics Consultant
Structure business problems, choose the right analysis and present findings to decision makers. First proof: a short case study deck from a real dataset.
Stage 3: where each data role leads
Common progressions after the first few years:
- Data Scientist → Senior Data Scientist → Lead or Principal Data Scientist.
- ML Engineer → Senior ML / MLOps Engineer → ML Platform Lead.
- Data Engineer → Senior Data Engineer → Data Architect.
- BI Analyst → BI Developer → BI Manager or Head of BI.
- Data Analyst → Senior Analyst → Product or Analytics Manager, or a move into data science.
- Analytics Consultant → Senior Consultant → Analytics Practice Lead; long term, Chief Data Officer.
Signs you are ready to leave the foundation stage
Before you branch, check that you can do these without looking things up constantly:
- Explain a p-value and a confidence interval in plain words.
- Write a SQL query with a join, a GROUP BY and a window function.
- Clean a messy CSV in Pandas and say what you changed and why.
- Produce three charts that each answer a different question about one dataset.
Which data branch suits your starting point
Where you start shapes which branch you reach fastest.
You enjoy maths and experiments
The Data Scientist branch rewards careful modelling and honest evaluation.
You like building systems that just work
Data Engineer or ML Engineer branches suit people who think in pipelines and deployments.
You come from commerce, business or operations
BI Analyst, Data Analyst or Analytics Consultant use your business sense from day one.
You are a fresher unsure which to pick
Start as an analyst-leaning learner. It is the easiest branch to move from into data science later.
Quick answers about the data science roadmap
Short answers to the roadmap questions data learners ask most.
What should I learn first in data science?
Statistics basics, Python with Pandas and SQL. Every data role, from analyst to ML engineer, depends on these three.
Is data engineering part of the data science roadmap?
It shares the same foundation, especially SQL and cleaning, then branches into pipelines and data modelling instead of machine learning.
Do I need machine learning to work in data?
Not for analyst or BI roles. Machine learning is essential for data scientist and ML engineer branches.
Can I start as a data analyst and become a data scientist?
Yes, it is one of the most common routes. You keep the SQL and analysis skills and add machine learning on top.
How is this different from the data analytics roadmap?
The data analytics roadmap goes deeper into Excel, BI, reporting and KPIs. This one adds machine learning, deployment and engineering branches.
Go deeper on each data science role
Open the role page for your branch, or read the guides that fill in each stage.
Test yourself against the foundation checklist
Go through the four "ready to branch" checks above with one public dataset. Whatever you cannot do yet is your next week of study; once all four are easy, pick a branch and start its first proof project.

