Before you pay for Python with data science training in Nepal, spend two weeks on free material. Learn plain Python first, then pandas in a browser notebook, then do a short test on your own numbers. If you enjoy the test, paid training will move you much faster. If you hate it, you have saved your money.

The test is the part most people skip. It is also the part that tells you the most.

Where should a complete beginner start with Python?

Start with Python for Everybody, not the official Python tutorial. The official tutorial on python.org says it is written for programmers who are new to Python, not for people who are new to programming. A beginner who opens it can get stuck early and decide they are "not a coding person". The problem is the tutorial, not the beginner.

Python for Everybody is free. Its welcome text says you can learn Python "even if you have no programming background". The lessons go in a sensible order: variables, conditions, functions, loops, strings, files, lists and dictionaries.

For data work, stop after the Dictionaries lesson. Network programming and object-oriented programming can wait until later.

Do you need a powerful laptop to start?

No. For the free part, use Google Colab. It runs Jupyter notebooks in your browser, and Google's FAQ says it is free and needs no setup. Your laptop only shows the page, so an old one is fine.

Two limits matter. Free notebooks run for at most 12 hours, and they stop when you leave them idle. Google also says free resources are not guaranteed. So save your notebook to Google Drive often, more so on days when the connection is weak.

Do not start by installing Python, editors and libraries on your own. Setup errors are a poor first lesson. They teach you nothing about data, and they make Python feel harder than it is. Install it locally later, when someone can sit with you and fix the path problems.

What should the second week look like?

Once loops and dictionaries feel normal, open 10 minutes to pandas. It is the official short introduction from the pandas team. pandas is the Python library analysts use to work with tables of data.

You do not need the whole page. Type out four sections in Colab: Object creation, Selection, Grouping and Plotting. In Selection, spend time on boolean indexing, like df[df["A"] > 0]. In Grouping, make sure you understand what groupby followed by sum() returns.

Type the code by hand, do not paste it. Pasted code runs, and you learn nothing from it. Typed code breaks in small ways, so you read the error and fix it. Reading an error message calmly is a skill every data job needs.

How do you know you are ready to pay for training?

Do this test with numbers you already know. Use your phone recharge spending for the last 12 months, or your family's monthly electricity bill. Type them into a Python dictionary in Colab, with one list for the months and one for the amounts.

Then do these steps, searching only for small bits of syntax:

  • Turn the dictionary into a pandas DataFrame.
  • Show only the months where you spent more than your average.
  • Add a column for the season, then use groupby to get the total per season.
  • Draw a line chart of all 12 months.
  • Write two plain sentences in a text cell about what the chart shows.

If you finish in an evening or two, you are ready. Paid training will then go into harder things than syntax, and you will get more from every class.

If you get stuck on the first step, go back to the Lists and Dictionaries lessons. That is not a failure. It shows you exactly what to practise next.

The last step matters most. A chart on its own answers nothing. The two sentences that say what it means are what a manager actually reads.

What can free material not give you?

Free material teaches syntax well. It is weak at feedback. Nobody tells you that your total is wrong because of a duplicate row, or that your chart hides the real story.

It also skips the messy parts of data in Nepal. Office records here often use Bikram Sambat dates, and turning them into normal dates is real cleaning work. Tutorials made abroad rarely mention it.

If you work in Excel or Google Sheets every day, you have a head start. A lookup like VLOOKUP does a job close to merge in pandas, which joins two tables on a shared column. If you know one, the other makes sense quickly.

How we teach Python for data science at Saarathi Academy

At Saarathi Academy in Old Baneshwor, Kathmandu, the Data Science and Machine Learning course runs for 12 weeks, in person and online. You do not need to know Python, statistics or maths before you join. Each one starts from zero when the course reaches it.

Python gets one and a half weeks on its own before any data library. pandas comes in Week 3, including turning Bikram Sambat dates into normal dates. SQL with PostgreSQL comes in Week 4. Each batch has up to 10 students in the classroom, and the full syllabus is public before anyone pays.

The free work above is not required, but it makes the first weeks easier. Our syllabus lists its own optional warm-up, including the first lessons of freeCodeCamp's Scientific Computing with Python and Khan Academy's Statistics and Probability.

For the course itself you need your own laptop with at least 8GB of RAM. Power BI Desktop, used in Week 8, runs only on Windows, so Mac users need a Windows setup before then.

Your next step tonight

Open Colab and start a new notebook. Type a dictionary with your last three months of phone recharge amounts, and turn it into a pandas DataFrame. That is step one of the test, and it takes about ten minutes.

When you want the full week-by-week plan, read the Data Science and Machine Learning course page and the syllabus.