loading…
Where it ends: Prepare a dataset with pandas, train a regression and a classifier with scikit-learn, and judge whether the model is any good.
The first 5 lessons are free, no account needed. The rest are in Pro, ₹299/month. Students and graduates who want the first rung of data science, with the maths kept to what the code needs.
Most of a data scientist's Python time goes on the data, not the model: loading, cleaning, joining and reshaping tables in pandas and NumPy, then plotting to see what is there. The modelling part on this path is scikit-learn: split the data, fit linear regression and a classifier, try a decision tree, and learn to read the evaluation numbers so you do not fool yourself with overfitting. Probability and statistics, deep learning and experiment design are part of the real job and are not taught here yet.
Take the lessons top to bottom. Each opens with code you run in the browser and ends with graded exercises. Already know the foundations? The level test tells you where to jump in.
After this: Read and write everyday Python: variables, loops, lists, dicts, functions, errors.
After this: Classes for your own types, JSON for other people's data.
After this: Arrays without loops, tables cleaned, grouped, merged and charted.
After this: Train, test and honestly evaluate your first models.
Free and printable. Every snippet on them was run on Python 3.12 before publishing.
Free. Write your answer and run the tests in the browser; the tests check correctness, not speed.
The one data project in the library today: pandas end to end on a real CSV. There is no model-training project yet; the ML lessons carry their own graded exercises.
Finishing this path gives you the Python part of the job. It is not, by itself, a qualification for it. The Data Scientist role page has the skills list and the workspace templates.