NPTEL and SWAYAM for data careers: choosing and showing courses
In this article (6 sections)
Related programme: the Data Science course (6 months; ₹1,30,000 (inclusive of GST) · EMI from ₹5,000/month).
NPTEL is a useful, low-cost way to add graded structure to self-study for a data career: pick a short sequence of courses in statistics, programming and machine learning, do the weekly assignments properly, and sit the proctored exam for the ones that matter most. The certificate helps as a signal of discipline; the skill only counts in hiring when you apply it in a project you can show.
How the programme works, from the official pages
Facts below are from NPTEL's site and its FAQ, checked on 27 September 2026. Rules can change each semester, so reread the course page before registering.
- NPTEL is an initiative of seven IITs (Bombay, Delhi, Guwahati, Kanpur, Kharagpur, Madras and Roorkee) and IISc, funded by the Ministry of Education. Its courses are also offered through SWAYAM.
- Certification courses run in two semesters a year (January–April and July–December) and last 4, 8 or 12 weeks.
- To be certified you join the course, submit weekly assignments, register and pay for the exam, and write an in-person proctored exam at a centre.
- For most courses the pass rule is: average assignment score of at least 40/100 and final exam score of at least 40/100. The final score weights assignments at 25% and the exam at 75%. Some courses use different rules, stated on their announcement page.
Exam fees and centre lists are published per semester on NPTEL's site. Check them there rather than relying on a figure someone quotes.
Work out your score before the exam
Because assignments are only a quarter of the final score, a strong assignment record cannot rescue a weak exam. This small calculation, with hypothetical scores, shows how the two combine:
def nptel_result(assignment_avg, exam_score):
"""Most-courses rule from the NPTEL FAQ: both components >= 40/100,
final = 25% assignments + 75% exam. Check your course page for exceptions."""
final = 0.25 * assignment_avg + 0.75 * exam_score
passed = assignment_avg >= 40 and exam_score >= 40
return round(final, 2), passed
for a, e in [(90, 38), (55, 62), (40, 40), (100, 70)]:
final, ok = nptel_result(a, e)
print(f"assignments {a:>3}, exam {e:>3} -> final {final:6.2f} {'PASS' if ok else 'FAIL'}")The first case, 90 in assignments and 38 in the exam, gives a final score of 51.0 but still fails, because the exam is below 40. Plan revision time for the exam, not just the weekly quizzes.
A sequence that builds a data skill
Choose courses by what they teach, not the title. Open the syllabus and the first week's lectures before registering. A sensible order for someone aiming at analyst or data science work:
| Order | Area | What to look for in the syllabus | Why this order |
|---|---|---|---|
| 1 | Programming in Python | Data types, functions, files, basic libraries | Every later course assumes it |
| 2 | Probability and statistics | Distributions, estimation, hypothesis tests, regression | Needed to interpret any model or A/B test |
| 3 | Databases or data management | Relational model, SQL queries | Most analyst tests include SQL |
| 4 | Introductory machine learning | Supervised learning, evaluation, overfitting | Build on statistics, not before it |
| 5 | Optional depth | Deep learning, NLP, optimisation, or a domain course | Only once 1–4 are solid |
One 8- or 12-week course per semester alongside college is realistic. Two is possible in a light semester. Three usually means none of them gets proper attention.
Get more than a certificate out of each course
For every course you take, produce one artifact outside the course:
- Python course: a script that cleans a messy CSV (inconsistent dates, duplicate rows) and writes a summary.
- Statistics course: a short write-up testing one claim on a public dataset, with the test, the assumptions and the result in plain words.
- Database course: ten SQL queries against a sample database, each answering a stated business question.
- ML course: one model with a proper validation split, a baseline and an honest paragraph on what it can't do.
Put these in a repository with a README naming the course that taught the method. That link, not the PDF certificate, is what an interviewer will open.
Show it on your resume and in interviews
Put certified courses under a "Certifications" heading, with the course name, the institute that offered it, the semester and your score if it is good:
"Probability and Statistics — NPTEL (offered by IIT [name]), Jul–Dec 2026, certified, score [x]%."
Then link the artifact under Projects. In interviews, expect: "What was the hardest topic, and how did you use it?" Answer with the project, not the lecture.
If you're earning credits through SWAYAM for your degree, confirm with your college's NPTEL or SWAYAM coordinator whether a particular course counts before you register; that depends on your university's rules.
When NPTEL isn't enough on its own
NPTEL courses are strong on concepts. They generally leave you to find the practice on messy data, the portfolio review and interview preparation. If you want those added with mentor feedback, the Data Science programme is six months of live, reviewed project work, taught online with optional Noida labs. Before deciding, compare data analytics and data science to see which direction your courses point to.
Frequently asked questions
Is an NPTEL certificate valuable for a data analyst or data science job?
It is a graded, proctored credential from a programme run by IITs and IISc, which makes it more credible than an ungraded online certificate. Employers will still test your skills directly, so treat the course as structured learning and a signal of discipline, then show a project that applies what you learned.
How do you pass an NPTEL course?
According to NPTEL's FAQ, for most courses you need an average assignment score of at least 40 out of 100 and a final exam score of at least 40 out of 100, with assignments weighted 25% and the proctored exam 75%. Some courses, such as those with non-proctored programming exams, use different criteria stated on the course page.
Pankit Kumar has 10 years in Data Science & AI, building and shipping production systems in regulated pharma and clinical environments. He is a freelance trainer at Boston Institute of Analytics, AnalytixLabs and Scaler, and has taught this material to thousands of working professionals.
This article is part of our Data Science programme — 6 months. From data foundations to machine learning, deep learning and deployment.
Explore Data Science