data[py]

SELECT * FROM celebs

Bootcamp (les plus chers):

IronHack (en français, 4 weeks, 7 500€)

Le Wagon (en français, 9 weeks, 6 900 €)

Lion (en français, 1 week, 1 499€)

Cours en ligne (les plus accessibles):

Codecademy (en anglais, gratuit)

Datacamp (en anglais, gratuit)

Dataquest (en anglais)

Openclassroom (en français, payant)

Le Wagon On-demand (en français, 150€)

Lion (en français, 1 499€)

Youtube, etc...

Cours en ligne (les plus techniques):

Coursera (Introduction to DS, en anglais, gratuit)

Coursera (Machine learning by Standford U., en anglais, gratuit)...

Parcours académiques (les plus sélectifs):

Data

Data analysis

Data mining

Data sciences

La Data Science mélange des outils, des algorithmes et des principes de machine learning avec pour objectif de découvrir des patterns cachés à partir de données brutes. Toute la différence entre les deux se résume à la différence qu’il y a entre expliquer et prédire.

Data visualization

Link: D3.js

Data wrangling

Régression linéaire (simple/multiple)

Pourquoi les Data Sciences ?

A Day in the Life - Data Analyst

It’s Catherine‘s first day working as a Data Analyst at Codecademy! She’s excited to use some of the tools that she learned in the Data Science Path, such as:

Data extraction with SQL

Programming basics with Python

Data analysis using pandas, a Python library

Data visualization using Matplotlib, a Python library

Machine Learning using scikit-learn, a Python library

Exploring Data with SQL

Like most organizations, Codecademy uses SQL (Structured Query Language) to access its database.

A database is a set of data stored in a computer. This data is usually structured into tables. Tables can grow large and have a multitude of columns and records.

Spreadsheets, like Microsoft Excel and Google Sheets, allow you to view and manipulate data directly: with selecting, filtering, sorting, etc. By applying a number of these operations you can obtain the subset of data you are seeking.

SQL (pronounced “S-Q-L” or “sequel”) allows you to write queries which define the subset of data you are seeking. Unlike Excel and Sheets, your computer and SQL will handle how to get the data; you can focus on what data you would like. You can save these queries, refine them, share them, and run them on different databases.

It is a great way to access data and a great entry point to programming because its syntax (the specific vocabulary that gives instructions to the computer) is very human-readable. Without knowing any SQL, you might still be able to guess what each command will do.

On her first day at Codecademy, Catherine wants to become familiar with the company’s data, so she connects to the database and uses SQL to explore the database.

Next, Catherine wants to take a look at the churn rate.

Churn rate is the percent of subscribers to a monthly service who have canceled. For example, in January, let’s say Codecademy had 1,000 learners. In February, 200 learners sign up, and 250 cancel.

The churn rate for February would be:

Catherine wants to analyze the churn rates for Codecademy for the past few months so she writes another SQL query.

Time for Python!

After interacting with the database, it is time to analyze the data further and eventually visualize the data. And SQL cannot get us there.

Python is a general-purpose programming language. It can do almost all of what other languages can do with comparable, or faster, speed. It is often chosen by Data Analysts and Data Scientists for prototyping, visualization, and execution of data analyses on datasets.

There’s an important question here. Plenty of other programming languages, like R, can be useful in the field of data science. Why are so many people choosing Python?

One major factor is Python’s versatility. There are over 125,000 third-party Python libraries. These libraries make Python more useful for specific purposes, from the traditional (e.g. web development, text processing) to the cutting edge (e.g. AI and machine learning). For example, a biologist might use the Biopython library to aid their work in genetic sequencing.

Additionally, Python has become a go-to language for data analysis. With data-focused libraries like pandas, NumPy, and Matplotlib, anyone familiar with Python’s syntax and rules can use it as a powerful tool to process, manipulate, and visualize data.

Analyzing Data with Pandas

Catherine wants to know if Codecademy learners are finding what they need. She needs to analyze data from a survey that was administered to a subset of visitors.

Catherine selects the survey data from the SQL database and downloads it onto her computer. She saves the data as a CSV file (extension .csv), which stands for Comma-Separated Values. CSV is a text-only spreadsheet format that lets us store and explore data.

To analyze the survey data, Catherine will use pandas. Python is a programming language, and pandas is a special set of commands in Python that lets us analyze spreadsheet data. Pandas can do a lot of the things that SQL can do, but it’s also backed by the power of Python, so we can easily transition from analyzing our data with pandas to visualizing it using other Python tools.

Visualizing Data with Matplotlib and Seaborn

Catherine wants to visualize her analysis and share it with her boss. For this, she will use Matplotlib, another Python module.

Matplotlib lets Catherine create line charts, bar charts, pie charts, and more.  It gives her precise control over colors and labels so that she can create the perfect chart to communicate her findings.

Catherine has written some code using Matplotlib that visualizes hours of usage on Codecademy!

Machine Learning

Machine Learning is the science of getting computers to learn and act like humans do, and improve their learning over time in autonomous fashion, by feeding them data and information in the form of observations and real-world interactions. It has taken the Data Science world by storm.

It can be branched out into the following categories:

Supervised Learning

Unsupervised Learning

Supervised Learning is when the data is labeled and the program learns to predict the output from the input data. For instance, a supervised learning algorithm for credit card fraud detection would take as input a set of recorded transactions. It would learn what makes a transaction likely to be fraudulent. Then, for each new transaction, the program would predict if it is fraudulent or not.

Unsupervised Learning is where the data is unlabeled and the program learns to recognize the inherent structure of the input data. For the same fraud example, the model would take in a bunch of transactions with no indication of if they are fraudulent or not, and it would group them based on patterns it sees. It might discover two groups, fraudulent and legitimate.

Begin Your Journey

Catherine learned and applied a lot during the first two weeks of her job. She:

Explored churn data using SQL

Analyzed data using pandas, a Python library

Visualized data using Matplotlib, a Python library

Built a Machine Learning model using scikit-learn, a Python library

Now it’s your turn. By the end of the Data Science Path, you’ll be able to do all of the things that Catherine can do. We will also be adding new content every month.