The Data Frog

Data Science and Machine Learning with python


I'm Colin, CNRS researcher, particle physicist at CERN, and CTO of Cynapps.

You'd like to get started with machine learning and data science?

You're at the right place.

No need for a university degree, or a strong background in programming. Just dive in!

Tim Manhalter


It rules! Learn just what you need for data science and machine learning.

python for machine learning

Python Crash Course for Machine Learning

You want to get started with machine learning but you don't know python? You're at the right place! (1h course)

numpy for machine learning

Numpy Crash Course for Machine Learning

Learn the minimum numpy needed to get started with machine learning (1h course)

matplotlib for machine learning

Matplotlib for Machine Learning

Learn the basics of matplotlib in 1h. You'll make your first plots with a toy machine learning example.

More Posts →

Machine Learning

Basics of machine learning and neural networks, unsupervised learning.

neural network with a single neuron

The 1-Neuron Network: Logistic Regression

The most simple neural network. Learn how a neuron is working.

non-linear classification problem

Logistic Regression vs Neural Network: Non Linearities

What are non-linearities and how hidden neural network layers handle them.

garbage in, garbage out in a neural network

Overfitting Illustrated

One of the most important issues in machine learning, illustrated in a small 2D classification problem.

More Posts →
h heyerlein

Deep learning

Build your first deep neural networks, and learn advanced techniques such as transfer learning.

pretrained transformers

Pre-trained Transformers with Hugging Face

Get started with the transformers package from Hugging Face for sentiment analysis, translation, zero-shot text classification, summarization, and named-entity recognition (English and French)

a dog cosplaying as a cat

Image Recognition: Dogs vs Cats! (92%)

Classify dog and cat pictures with a 92% accuracy with a deep convolutional neural network.

two robots transfer learning

Image Recognition with Transfer Learning (98.5%)

Use transfer learning to easily classify dog and cat pictures with a 98.5% accuracy.

More Posts →

Data Science

Statistics, data preparation, databases, you name it

colin working on a remote jupyter notebook

Remote jupyter notebooks with ssh port forwarding

Access your jupyter notebook server remotely

fit to the COVID-19 data and extrapolation, with uncertainties

COVID-19 Analysis: Uncertainties

Learn how to analyse the COVID-19 data from JHU by yourself, with a proper treatment of the uncertainties.

connecting to mongoDB from a jupyter notebook

MongoDB in 10' : Fill with python, Read with pandas

Set up your first MongoDB server, store data with python, and analyze it with pandas in a jupyter notebook.

More Posts →
Lucas Clara


Visualize your data before and after machine learning

visualizing datasets with t-SNE

Visualizing Datasets

Study variable correlations with matplotlib and seaborn, and use dimensionality reduction (PCA, t-SNE) to display complex datasets.

real estate transactions displayed on a google map

Show your Data in a Google Map with Python

Create an interactive display for geographical data with python: real-estate prices near Geneva.

choropleth map of the french population, by department

Choropleth Maps in Python (2021)

Create a choropleth map with geoviews and geopandas. Working geoviews installation instructions as of May 2021.

More Posts →

Accelerated Computing

Python is slow? Nope.

accelerating python with numba

Make python fast with numba

Python is an interpreted language, so it's flexible and easy to use, but it can be slow. Learn how to make it 100 times faster by compiling it for your machine, with just one line of additional code. Notebook ready to run on the Google Colab platform

boosting python with numba and cuda

Boost python with your GPU (numba+CUDA)

Use python to drive your GPU with CUDA for accelerated, parallel computing. Notebook ready to run on the Google Colab platform

GPU tu102 block diagram

CUDA kernels in python

Write your own CUDA kernels in python to accelerate your computing on the GPU. Notebook ready to run on the Google Colab platform

Mathew Schwartz

Internet Of Things

Raspberry pis, microcontrolers, sensors, and how to communicate with them

Home assistant dashboard

Open source smart home with home assistant and docker

Deploy home assistant with docker on a raspberry pi: step-by-step tutorial

raspberry pi zero W

Raspberry pi Zero W : Headless Install

Cute little computer with Wifi access. Easy recipe for a headless install of raspbian lite, without screen and keyboard.

data pipeline from sensors to database to visualization

Data pipeline with Docker, InfluxDB, and Grafana

Set up a typical data pipeline as may be used in experimental science and IOT, with Docker, InfluxDB, and Grafana.