Python, machine learning and analytics
Data science tutorials you can follow from start to finish.
Step-by-step guides built on code we run ourselves, with the real outputs, so you get the same results on your own machine.
Popular: pandas, SQL, XGBoost, TensorFlow, Power BI
342 real penguins, measured three ways
Drag to turn the chart. HoverTap a point for details.
New to data science? Start here
Four tutorials, in this order, take you from installing Python to your first model.
- Install Python and Jupyter Notebook
Set up everything you need in one sitting.
- Data manipulation with pandas
Load, filter, group and reshape tables.
- Clean and preprocess data
Missing values, duplicates and outliers.
- Your first model with scikit-learn
Train, test and evaluate a model.
Recently rewritten
Every one rebuilt around code we ran, with the real results.
How to deploy a machine learning model with Flask
Save the pipeline, build a validated API, test it and serve it with Gunicorn and Docker.
TensorFlow for beginners: build your first neural network
Tensors, gradients and a Keras network that predicts diamond prices.
K-means clustering explained, with a Python example
How k-means works, why scaling matters and when to use DBSCAN instead.
Random forests explained, with a Python example
Bagging, feature importance and an honest comparison on the Titanic data.
Convolutional neural networks explained, with Keras
Filters, pooling and a CNN that reaches 91% on Fashion-MNIST.
Principal component analysis (PCA) in Python
What PCA does, how many components to keep and how to use it in a pipeline.
Free data tools
Converters and calculators that run in your browser. Your data is never uploaded.
Learning paths
Pick a topic and work through it in order.
Python for data analysis
Beginner to intermediate
Machine learning
Intermediate
Deep learning
All levels
Data science by industry
How companies in different fields put data to work.
About Pulse Data Hub
We write practical tutorials on Python, machine learning and analytics. Every tutorial is built around code we run before publishing, so you can follow along and get the same results.
More about us →Get new tutorials by email
We send new guides when they are published. Unsubscribe anytime.
