onarcle

Artificial intelligence

Data science and machine learning with Python

From raw data to predictive models

Five days to prepare and explore data, build and evaluate machine learning models with Python (pandas, scikit-learn) and present results that support decision-making.

Duration
5 days · 35 hours
Level
Intermediate
Format
In-company · Virtual classroom

Upcoming sessions

This course is run on demand: in a virtual classroom or at your premises, on dates that suit you. Ask us for the next dates or a quote for your team.

Also available

  • In-company, at your premises: France
  • In a virtual classroom: France

Overview

This intensive course takes you through every step of a data science project on real datasets: preparation, visual exploration, modelling, evaluation and reporting. You learn to choose the right algorithm, avoid classic pitfalls and explain your results to non-specialists.

Objectives

  • Prepare and clean data with pandas
  • Explore and visualise data
  • Build regression and classification models
  • Evaluate and improve a model
  • Present results and prepare a model for production

Who is it for?

  • Analysts, management controllers, research officers
  • Developers moving into data
  • Engineers and scientists

Prerequisites

Basic Python (variables, loops, functions) and descriptive statistics. The "Python fundamentals" course is recommended.

Programme

  1. Day 1

    Preparing data

    • Python environment and notebooks
    • pandas: loading, filtering, transforming
    • Missing values and outliers
  2. Day 2

    Exploring and visualising

    • Descriptive statistics
    • Visualisation with Matplotlib and Seaborn
    • Formulating hypotheses
  3. Day 3

    Supervised learning

    • Linear and logistic regression
    • Decision trees and random forests
    • Train / test split
  4. Day 4

    Evaluating and improving

    • Metrics: precision, recall, AUC, error
    • Cross-validation and hyperparameter tuning
    • Unsupervised learning: clustering
  5. Day 5

    End-to-end project

    • Project on a real case
    • Interpreting and explaining a model
    • Reporting and moving to production

Teaching method

Guided notebooks and a running project on real data. Groups of up to 10 participants, in a virtual classroom or at your premises, with a practitioner trainer. For 30 days after the course, you can email your questions to the trainer.

What you receive

  • Annotated Python notebooks for every exercise
  • pandas and scikit-learn cheat sheets
  • Data science project methodology

Assessment and certificate

Initial assessment, hands-on exercises and a final quiz. Certificate of attendance; Onarcle certificate of achievement from a 70% score on the final quiz.

Price and funding

Public course price per participant, materials included. In-company (up to 10 participants, programme tailored to your needs): €6,990 excl. VAT per group. 10% discount from 3 participants from the same organisation. If your employer or a funding body will pay, contact us and we will find the right arrangement.

Enrolment and quotes

Book a place, request a quote or a session for your teams. We reply within 2 working days.

Software and brand names mentioned belong to their respective owners. Onarcle is an independent training provider, not affiliated with their publishers; this course does not award an official vendor certification.

← View the full catalogue