Python, Visually

Learn Python by running it. Every guide is a set of notebooks that open in Google Colab: read a little, run a cell, change it, run it again. Nothing to install and nothing to set up.

New to Python? Start with Python from the Start. The guides below are in reading order.

93guides

250notebooks planned

172written so far

0things to install

How a notebook works. A notebook opens in Google Colab and runs in your browser. It contains the explanation, worked examples you can edit, exercises to try, and a section on the errors that part of Python actually produces, with the real message shown.

The exercises are the point. Every notebook has a separate solutions file, so the answer is never sitting one scroll below the question.

The guides

Foundations teaches the language itself, and is best worked through in order: each guide depends on the one before it. The deep dives take one library apiece, grouped by what the library is for, and any of them can be read once you have the foundations it names. Projects puts everything to work on real problems.

Foundations
1	Python from the StartThe language itself	21	Complete
2	Files, Paths and FormatsGetting data in and out	10	Complete
3	Object-Oriented PythonClasses, and when not to use them	14	Complete
4	APIs and JSONTalking to the rest of the internet	18	Complete
5	Testing and PackagingMaking it survive contact with other people	10	Complete
Relational and Document Databases
6	SQLAlchemy, Deep DiveSessions, models and the SQL underneath	20	Complete
7	SQLModel, Deep DiveOne model for the table and the API	14	Complete
8	Peewee, Deep DiveA small mapper, end to end	12	Complete
9	sqlite3, Deep DiveThe database already on your machine	20	Complete
10	asyncpg and psycopg3, Deep DivePostgreSQL, from Python	17	Complete
11	PyMongo and Beanie, Deep DiveDocuments instead of rows	16	Complete
Data Science and Machine Learning
12	NumPy, Deep DiveArrays, shapes and vectorization	14	Planned
13	Pandas, Deep DiveTabular data, properly	18	Planned
14	scikit-learn, Deep DiveEstimators, pipelines, evaluation	14	Planned
15	PyTorch and TensorFlow, Deep DiveTensors, gradients and training loops		Planned
16	LightGBM and XGBoost, Deep DiveGradient boosting on tables		Planned
17	GeoPandas, Deep DiveTables with geometry in them		Planned
18	Shapely, Deep DivePoints, lines and polygons		Planned
Data Engineering and Analytics
19	Polars, Deep DiveExpressions, and the query it plans		Planned
20	DuckDB, Deep DiveSQL over files, with no server		Planned
21	PyArrow, Deep DiveThe columnar format underneath		Planned
22	PySpark, Deep DiveWhen the data outgrows the machine		Planned
Statistics and Econometrics
23	SciPy, Deep DiveOptimization, signals and statistics	12	Planned
24	Statsmodels, Deep DiveRegression with a summary table		Planned
25	PyMC, Deep DiveBayesian models you can sample		Planned
26	Pingouin, Deep DiveTests that print a table		Planned
AI Agents and Retrieval
27	LlamaIndex, Deep DiveYour documents, made retrievable		Planned
28	Haystack, Deep DiveSearch pipelines that reach production		Planned
29	DSPy, Deep DivePrompts written by a program		Planned
30	LangChain and LangGraph, Deep DiveTools, state and the loop		Planned
31	CrewAI, Deep DiveAgents with roles		Planned
32	AutoGen, Deep DiveAgents that talk, and run code		Planned
33	PydanticAI, Deep DiveAnswers your code can trust		Planned
34	smolagents, Deep DiveAgents that write Python		Planned
LLM Training, Fine-Tuning and Serving
35	Unsloth, Deep DiveFine-tuning that fits in your memory		Planned
36	Axolotl, Deep DiveA training run from one file		Planned
37	Torchtune, Deep DiveRecipes you can read		Planned
38	LLaMA-Factory, Deep DiveMany models, many methods		Planned
39	PEFT and TRL, Deep DiveAdapters, and training on preferences		Planned
40	Accelerate, Deep DiveOne script, several cards		Planned
41	DeepSpeed and Megatron-LM, Deep DiveTraining that will not fit		Planned
42	PyTorch FSDP, Deep DiveSharding, in PyTorch itself		Planned
43	vLLM, Deep DiveServing a model to many people		Planned
44	llama-cpp-python and Ollama, Deep DiveModels on your own machine		Planned
Computer Vision and Image Processing
45	OpenCV, Deep DiveFrames, filters and features		Planned
46	Ultralytics YOLO, Deep DiveDetection, segmentation and tracking		Planned
47	MediaPipe, Deep DiveFaces, hands and pose, in real time		Planned
48	Pillow and Albumentations, Deep DiveLoading images, and changing them on purpose		Planned
Voice, Speech and Audio
49	Whisper and faster-whisper, Deep DiveSpeech to text, offline		Planned
50	Vosk, Deep DiveRecognition on a small machine		Planned
51	SpeechRecognition, Deep DiveOne interface, several engines		Planned
52	Coqui TTS and XTTS, Deep DiveSynthetic speech, and cloned voices		Planned
53	edge-tts, Deep DiveHundreds of voices, over the network		Planned
54	pyttsx3, Deep DiveThe voices already installed		Planned
55	Librosa, Deep DiveWhat a sound looks like		Planned
56	Pydub and PyAudio, Deep DiveEditing audio, and recording it		Planned
Game Development
57	Pygame and Pygame CE, Deep DiveThe loop, the sprite and the clock		Planned
58	Arcade, Deep Dive2D with the graphics card doing the work		Planned
59	Pyglet, Deep DiveWindows, graphics and no dependencies		Planned
60	Ursina, Deep Dive3D, quickly		Planned
61	Panda3D, Deep DiveA 3D pipeline you control		Planned
62	UPBGE, Deep DiveGames inside Blender		Planned
63	Ren'Py, Deep DiveBranching stories, and saves		Planned
64	Pymunk, Deep DivePhysics you can trust		Planned
Web Frameworks and Networking
65	FastAPI, Deep DiveAsynchronous APIs, typed and documented		Planned
66	Django, Deep DiveThe framework that brings everything		Planned
67	Flask, Deep DiveSmall, until you decide otherwise		Planned
68	Requests and httpx, Deep DiveSessions, timeouts and retries		Planned
69	BeautifulSoup and Scrapy, Deep DiveReading a page, and crawling a site		Planned
Interactive Apps and Visualization
70	Matplotlib, Deep DivePlots you can defend	12	Planned
71	Seaborn, Deep DiveStatistical plots in one line		Planned
72	Plotly, Deep DiveCharts that answer back		Planned
73	Streamlit, Deep DiveA script that becomes an app		Planned
74	Gradio, Deep DiveA demo somebody else can try		Planned
Workflow Orchestration and Background Work
75	Dagster and Prefect, Deep DivePipelines that know their dependencies		Planned
76	Apache Airflow, Deep DiveSchedules, and filling in the past		Planned
77	Celery, Deep DiveWork that happens later		Planned
Developer Tooling and Quality
78	pytest, Deep DivePlugins, and testing hard things		Planned
79	Pydantic, Deep DiveTypes checked while the program runs		Planned
80	uv and Ruff, Deep DiveFast tools for the dull parts		Planned
Finance and Quantitative Analysis
81	yfinance, Deep DiveMarket data, into a DataFrame		Planned
82	Backtrader, Deep DiveA strategy, tested against the past		Planned
83	TA-Lib, Deep DiveIndicators, computed correctly		Planned
84	CCXT, Deep DiveOne client, many exchanges		Planned
Documents and File Formats
85	pypdf and PyMuPDF, Deep DiveText out of a PDF, and pages back in		Planned
86	openpyxl and XlsxWriter, Deep DiveSpreadsheets, written properly		Planned
87	python-docx, Deep DiveWord documents from code		Planned
Parallel Computing and Concurrency
88	Ray, Deep DivePython across a cluster		Planned
89	Dask, Deep DiveNumPy and Pandas, in parallel		Planned
90	Joblib, Deep DiveA parallel loop, and a cache		Planned
Security and Networking
91	Scapy, Deep DivePackets, built and read		Planned
92	Impacket, Deep DiveProtocols at the level attackers use		Planned
Projects
93	Projects, End to EndEverything at once	8	Planned