Lumen AcademyBrowse courses

Applied Data Analysis with Python

Take a messy CSV from an actual business, clean it, question it, and hand a decision-maker something they can act on.

  • Intermediate
  • 46 hours over 8–12 weeks
  • Cohort + mentor, cohorts of 12
  • 4.7 (1,284 reviews)
  • 4,310 enrolled
  • Video
  • Reading
  • Practice
  • Project
  • Assessment

8 modules · 39 lessons · 46h of scheduled work

The whole syllabus

Every lesson, with the type of work it is and how long it is scheduled to take. Live mentor calls are marked as video because that is what they are; the two-week project is the only block with no fixed end.

  1. 1Getting your data in one piece

    5 lessons · 5h

    You start with the same 41 MB of branch exports everyone else gets, including the two broken files.

    • What you are given, and what is wrong with itVideo25 min
    • Reading CSV, Excel and a bad exportVideo40 min
    • Set up: Python, uv, JupyterReading30 min
    • Load all 34 branch files into one framePractice120 min
    • Checkpoint: row counts must matchPractice85 min

  2. 2pandas without the guesswork

    6 lessons · 7h

    The twelve operations that cover most analysis work, drilled until you stop looking them up.

    • Selecting: loc, iloc and the SettingWithCopy warningVideo45 min
    • groupby, agg and the shape it returnsVideo45 min
    • Merging frames and the join you got wrongVideo40 min
    • Reshaping: pivot, melt, stackPractice90 min
    • Drill set: 24 small tasksPractice180 min
    • Reference sheet you keepReading20 min

  3. 3Cleaning: nulls, duplicates, dates

    5 lessons · 6h

    Where the real hours go. Includes three date formats from the same source system.

    • Deciding what a null means hereReading30 min
    • Duplicate orders, and the ones that are notVideo40 min
    • Parsing dates that disagreePractice100 min
    • Outliers: keep, cap or dropVideo40 min
    • Write a cleaning script you can rerunPractice150 min

  4. 4Asking a question the data can answer

    4 lessons · 4h

    The module people skip and then regret. Mostly writing, no code.

    • Turning ‘why are sales down’ into three questionsReading60 min
    • What this dataset cannot tell youReading45 min
    • Write your question, get it reviewedPractice105 min
    • Mentor call: 30 minutes, one-to-oneVideo30 min

  5. 5Charts that survive a meeting

    5 lessons · 6h

    matplotlib and a house style. You will redraw the same chart four times.

    • Choosing the chart for the claimVideo40 min
    • matplotlib, enough of itVideo55 min
    • Labels, units and the axis that liesVideo35 min
    • Redraw: four versions of one chartPractice150 min
    • Critique thread with your cohortPractice80 min

  6. 6Statistics you will actually use

    7 lessons · 7h

    Six ideas, no proofs. Enough to stop you claiming something you cannot support.

    • Averages, medians and when to say neitherReading35 min
    • Variation, and describing it honestlyVideo45 min
    • Confidence intervals in plain wordsVideo50 min
    • Correlation, and the sentence you must not writeReading35 min
    • Simple A/B comparisonPractice120 min
    • Seasonality, spotted not modelledPractice90 min
    • Self-check: 15 questionsAssessment45 min

  7. 7The demand report

    4 lessons · 8h

    Two weeks. You write it for a regional head who has eleven minutes.

    • The brief, from the clientReading30 min
    • Draft one: analysis notebookProject180 min
    • Draft two: the four-page reportProject180 min
    • Mentor review and rewriteProject90 min

  8. 8Assessment and review

    3 lessons · 3h

    Graded against a rubric you can read before you start. Two attempts included.

    • The rubric, publishedReading15 min
    • Timed analysis task, new datasetAssessment120 min
    • Report defence, 20 minutes liveAssessment45 min

What you leave with

  1. 1

    A cleaning script that reruns

    One Python file that takes 34 raw branch exports and returns a single clean table. It is yours; three past learners now run it at work every Monday.

  2. 2

    A four-page demand report

    Written for a regional head with eleven minutes. Charts, a recommendation, and a section that says what the data cannot tell you.

  3. 3

    A marked notebook with the rubric attached

    Your analysis, with the instructor's line comments and your scores against the published rubric. The certificate links to it.

Before you enrol

  • You can write a Python loop and a function without looking them up. If you cannot, do the free 4-hour primer first — we will tell you on day one, not in week three.
  • You have used a spreadsheet for real work. Pivot tables help.
  • A laptop with 8 GB of RAM. The dataset is 41 MB; nothing here needs a GPU.
  • Roughly 5 hours a week for 9 weeks. The project fortnight needs 8.

Assessment and certificate

  • A 15-question self-check in module 6, unmarked, so you know where you stand.
  • A timed 2-hour analysis task on a dataset you have not seen, marked against a rubric published on this page.
  • A 20-minute live defence of your report with the instructor. Two attempts are included.
  • The certificate carries your project title, your four rubric scores and a link to the notebook itself, so it can be read rather than trusted.

How far people actually get

Share of the 4,310 enrolled learners who have opened each module. Half of the drop happens in module 3, which is the cleaning work, and we have rewritten it twice because of that.

  1. 100%1
  2. 92%2
  3. 81%3
  4. 74%4
  5. 68%5
  6. 61%6
  7. 54%7
  8. 49%8

49% reach the assessment · 44% submit the project · 41% are certified · median 9 weeks

Questions people ask first

What happens if I fall behind the cohort?
You keep access and join the next cohort's live hours at no extra cost. Roughly one in five learners does this; it is normal and you do not need to ask permission.
Is the certificate worth anything?
It is worth what your assessed work is worth. The certificate carries your project title, your rubric scores and a link to the work itself, so a hiring manager can read it rather than trust it.
Can I pay in instalments?
Yes. Any course above ₹4,999 can be paid over 3, 6 or 9 months at no extra cost through Razorpay, on most Indian credit cards and on cardless EMI from HDFC, ICICI, Kotak and Axis.
What if the course is not what I expected?
Full refund inside 14 days of your enrolment date, even if you have finished half of it. You keep nothing — access ends — but you keep your money. We refunded 213 learners last year.

People who took this one also took

Data & analyticsBeginner

SQL for Analysts

Joins, windows and CTEs, taught on a 2.1 million row order table you can break as often as you like.

6 modules · 22 h · Self-paced4.8(2,076)

Project: Rebuild a weekly revenue report as one query

₹3,499
One payment, lifetime access
See the syllabus
Data & analyticsIntermediate

Dashboards That Get Decisions Made

Most dashboards are never opened twice. This one is about the five numbers a team will actually look at on a Monday.

5 modules · 18 h · Self-paced4.4(613)

Project: A one-screen operations board reviewed by the instructor

₹4,299
One payment, lifetime access
See the syllabus
CommunicationIntermediate

Presenting Numbers to People Who Hate Numbers

Ten slides, one argument, no chart junk. You record yourself twice and watch it back, which is the whole course really.

4 modules · 12 h · Self-paced4.7(508)

Project: A ten-slide recorded presentation, reviewed

₹3,299
One payment, lifetime access
See the syllabus