Python for data science means using Python to load, clean, reshape and summarise data, usually in a notebook. We teach it from zero inside our Data Science and Machine Learning course. Plain Python gets a full week and a half before any data library. Then come NumPy and pandas, and then SQL.
What it is, and why it matters
Most data jobs start with messy data. A sales file has missing values. A date column is stored as text. Two branches spell the same product two ways. Python lets you fix all of that in code, so you can run the same steps again next month in seconds instead of redoing them by hand in Excel.
Three tools do most of this work:
- Jupyter notebooks: a page where you run code one block (cell) at a time and see the result right under it.
- NumPy: a library for fast maths on whole columns of numbers at once.
- pandas: a library for tables of data, much like a spreadsheet you control with code.
Python also sits under everything that comes later: charts, statistics, dashboards in Streamlit and machine learning in scikit-learn. If your Python is weak, every later topic feels hard. That is why we spend real time on it first.
What you learn, week by week
Week 1: Python basics
- Setting up Python, Jupyter, VS Code, Git and a first GitHub commit, together in class
- Variables, numbers, text and operators
- Decisions with if and else, and loops with for and while
- Functions: inputs, return values and scope
- Lists, dictionaries, tuples and sets
- A bridge from spreadsheets: SUM, IF, VLOOKUP and COUNTIF written in Python
- Notebook habits: clean cell order and Restart-and-Run-All, plus using ChatGPT and Copilot to explain code, never to run code you have not read
Week 2: more Python, then NumPy
- List and dictionary comprehensions (short ways to build lists)
- Reading and writing files with Python's standard library, including CSV files
- Handling errors and reading tracebacks (Python's error messages)
- Reusable modules, and finding a bug step by step
- A mini project: a CSV analyser in plain Python, with no library
- NumPy arrays: maths on a whole column at once, filters with boolean masks, and why arrays beat lists for numbers
Week 3: pandas and data cleaning
- Series and DataFrames, and the difference between .loc and .iloc
- Reading CSV and Excel files
- Pivot tables and lookups in Excel, then the same work in pandas with groupby and pivot_table
- Filtering, sorting, grouping and totals
- Missing values, duplicates and wrong data types
- Cleaning text and dates
- Joining tables with merge and concat, and reshaping with melt and pivot
- A full clean of a messy dataset, using Claude and Copilot as helpers and checking what they write
Week 4: Python meets SQL
SQL is the language for asking questions of a database. In Week 4 you learn it in PostgreSQL, then connect it to Python with read_sql and SQLAlchemy. You also write parameterised queries, which keep you safe from SQL injection (a common attack). The week ends with one question answered from four sources, CSV, Excel, a JSON API and Postgres, and the results matched up.
What you build and practise
Every class is hands-on. You type the code, it breaks, and you fix it. In Week 2 you build the plain-Python CSV analyser. In Week 3 you clean a messy dataset from start to finish in pandas. You use these skills for the rest of the course, through statistics, Power BI and machine learning, up to a final project that a classmate re-runs from a clean copy of your GitHub repo.
If you are wondering how long Python takes on your own, we wrote an honest answer in how long it takes to learn Python for data analysis. For a wider view of the field, read our data science with Python guide for 2026.
Who this suits
- Complete beginners. You do not need to know Python, statistics or maths.
- People who work in Excel all day and want to move to code.
- Fresh graduates from any subject who want an analyst job.
- CS or IT students who know some Python but have never used pandas on real data.
You need a laptop with at least 8GB of RAM and basic English reading. You do not install anything beforehand.
Taught inside the full course
Python for data is the first four weeks of the 12-week Data Science and Machine Learning course. It is not a separate class. After it come statistics, charts, Power BI, machine learning and a final project. Classes run Monday to Friday, 2 hours a day, in Old Baneshwor, with at most 10 students. Students outside Kathmandu can take a live online seat.
Common questions
Do I need to know Python before joining?
No. Python starts from zero in Week 1, and we set it up with you in the first class. Basic computer skills and basic English reading are enough.
How long do you spend on plain Python before pandas?
About a week and a half. Week 1 and the first half of Week 2 are plain Python. NumPy comes in the second half of Week 2 and pandas in Week 3.
Will I use AI tools to write code?
Yes, but carefully. You use ChatGPT, Claude and Copilot to explain and draft code, and you read and check everything before you run it.
Is Excel covered too?
Yes, alongside Python. Week 1 maps common Excel formulas to Python, and Week 3 covers pivot tables and XLOOKUP, then does the same work in pandas.