Python beginners
You know variables and loops, but arrays, DataFrames, and model pipelines still feel like separate worlds. Start with the orientation and build one mental model at a time.
Interactive Python · zero setup
NumPy, pandas, Matplotlib, and scikit-learn — taught with real code you can edit and run. No installs, no accounts, nothing to configure.
import numpy as np
x = np.arange(5)
squares = x ** 2
print(squares)
print(squares.mean())Output
[ 0 1 4 9 16]
6.0Real Python. Running locally.
The curriculum
Build a mental model for ndarrays, broadcasting, and vectorized thinking.
Load messy data, reshape it, and answer questions with groupby and joins.
Turn numbers into clear figures — then make them look like you meant it.
Fit models with the estimator API and evaluate them without fooling yourself.
How it works
Reading about an API is not the same as using it. Every lesson ends with code you write yourself, checked by tests that run where the code runs.
Each idea gets a page, a mental model, and a runnable example — no 40-minute videos.
Python loads in a Web Worker. Edit code, press Run, and see genuine library output.
Auto-checked tasks and quick quizzes tell you when an idea has actually landed.
Powered by Pyodide · no install · nothing leaves your device
Who it is for
LearnPY is a guided practice space, not a video library. Every track starts from a concrete mental model, then asks you to use it in a runnable example.
You know variables and loops, but arrays, DataFrames, and model pipelines still feel like separate worlds. Start with the orientation and build one mental model at a time.
You already ask good questions of data. Learn the pandas and Matplotlib patterns that make those questions repeatable instead of manual.
You have watched the tutorials. Now fit, evaluate, and debug real scikit-learn estimators until the workflow sticks.
Questions
No. LearnPY runs Python and the data libraries in your browser through Pyodide, so lessons, practice challenges, and the playground work without a local install or a notebook server.
Yes. The lessons, function reference, ML API atlas, practice challenges, playground, and capstone are all free to use.
The curriculum covers NumPy for arrays and vectorized math, pandas for tabular data, Matplotlib for visualization, and scikit-learn for machine learning. The reference includes 40 focused function pages and a 211-entry ML API atlas.
Code runs locally in your browser. Lesson progress, practice state, and playground solutions are stored on your device, and LearnPY does not require an account or sign-in.
Lessons, references, and progress pages are responsive and work in modern mobile browsers. The code editor is most comfortable on a larger screen, so practice sessions are easier on a laptop or desktop.
Ready when you are
Start with a two-minute orientation, or jump straight into NumPy. Progress is saved on this device.