Empire Training - Knwoledge forges empires

Data - Data Scientist

A data scientist analyzes and interprets complex data to aid in strategic decision-making and problem-solving.

Preview
Projects
Accompaniement

Objectives of Data Scientist training

Operational objective:

Know how to understand Data Scientist.

Educational objectives:

More concretely, at the end of this Data Scientist Fundamentals training you will have acquired the knowledge and skills necessary to:

Discover the job of Data Scientist and the main families of problems
Know how to model a Data Science problem
Create your first variables
Building Your Data Scientist Toolbox
Participate in a first competition.

Who is this training for?

Audience :

This internship is aimed at Analysts, Statisticians, Architects, Developers.

Prerequisites:

To follow this course in the best possible conditions, you need to have some basic knowledge of programming or scripting, as well as some memory of statistics which can be a plus.

A pedagogy based on practice

Acquire essential skills by validating professional projects.
Progress with the help of a professional expert.
Gain real know-how as well as a portfolio to demonstrate it.

Data Scientist Course Content:

Introduction to Big Data:

What is Big Data?
The Big Data Technology Ecosystem

Introduction to Data Science, the job of Data Scientist:

The vocabulary of a Data Science problem
From statistical analysis to machine learning
Overview of the possibilities of machine learning

Modeling a problem:

Input / output of a machine learning problem

“OCR” Practical Work:

How to model the optical character recognition problem.

Identify machine learning algorithm families:

Supervised analysis
Unsupervised analysis
Classification / regression

Under the hood of algorithms: linear regression:

Some reminders: hypothesis function, convex function, optimization
Construction of the cost function
Minimization method: gradient descent

Under the hood of algorithms: logistic regression:

Decision boundary
Construction of a convex cost function for classification

The Data Scientist's Toolbox:

Introduction to tools
Introduction to Python, Pandas and Scikit-learn

Practical case n°1: “Predicting Titanic survivors”

Statement of the problem
First manipulation in Python

The pitfalls of machine learning

Overfitting or overlearning
Bias vs. variance
Regularization: Ridge and Lasso regression

Data Cleaning

Data types: categorical, continuous, ordered, temporal
Detection of statistical outliers, aberrant values
Strategy for missing values

Practical work:

“Fill in missing values”

Feature Engineering

Strategies for non-continuous variables
Detect and create discriminant variables

Practical case n°2: “Predicting Titanic survivors”

Identifying and creating the right variables
Creation of a first model
Submission on Kaggle

Data visualization

Visualization to understand data: histogram, scatter plot, etc.
Visualization to understand algorithms: train/test loss, feature importance, etc.

Introduction to set methods

The basic model: the decision tree, its advantages and its limits
Presentation of the different ensemble strategies: bagging, boosting, etc.

Practical work “Return to the Titanic”:

Using a set method based on the previous model

Semi-supervised learning

The major classes of unsupervised algorithms: clustering, PCA, etc.

Practical work “Detecting anomalies in betting”:

How does an unsupervised algorithm detect fraud in betting?

Individual and privileged supervision.

The Empire Training community

Online pre-registration

Please fill out the form

How does an Empire Training course work?

From the chosen training to their entry into their new career, our students recount each stage of their experience and the support they received.