Machine Learning and Data Analytics

From descriptive analysis to predictive analysis

Objectives

At the end of the course the trainees will be able to recognize and use the main predictive algorithms and knowledge extraction from the data. Data import and processing tools will be used in a cloud environment, using Python language basics.

This training is aimed at various types of professionals who seek to evolve from descriptive analysis to predictive analysis using machine learning and data knowledge extraction techniques.At the end of the course the trainees will be able to recognize and use the main predictive algorithms and knowledge extraction from the data. Data import and processing tools will be used in a cloud environment, using Python language basics.

This course is aimed at various types of professionals who seek to evolve from descriptive analysis to predictive analysis using machine learning and data knowledge extraction techniques.

A professional training certificate will be issued on the SIGO platform, in accordance with Ordinance No. 474/2010 of July 8.

  • Who should attend

    Managers, analysts, controllers, CFOs, consultants, accountants or early career professionals looking to specialize in these areas

  • Resources

    Trainees must bring their laptop and should have a free google account, since Google Colaboratory will be used in this course

  • Requirements

    It's not mandatory but it would be beneficial to have basic knowledge of statistics and data analysis.

  • Duration

    16 hours

Programme Outline

  • Day 1

    Introduction

    • Artificial Intelligence (AI) framework in companies and businesses
    • Abstraction, generalization and evaluation

    Development Enviroment:

    • Google Colaboratory
    • Python essencial:
      • Pandas, Numpy, Scikit Learn, Matplotlib, Pycaret

    Exploratory Analysis

    • Measures of central trend and dispersion
    • Graphics:
      • Boxplots, histograms, scatterplots and correlation matrices

    Data Preprocessing

    • Data processing and cleaning
    • Feature engineering
  • Day 2

    Supervised Algorithms

    • Linear regression: how to develop a sales forecast model
    • Logistic regression: predicting customer abandonment
    • Decision trees: predict the class of wines and computer sales
    • Support vector machines: linear and nonlinear classification models
    • K-nearest neighbours: cancer detection model

    Unsupervised Algorithms

    • K-Means clustering: Segment customers according to their behavior
    • HDBSCAN: develop density-based cluster models
    • Association rules: identify which products sell well together

    Development and Improvement of Models

    • Training, validation, and testing data
    • Underfitting and overfitting
    • Regularization of models
    • Model performance evaluation
    • Hyperparameter tuning

    Practical Cases

    • AutoML: An automated machine learning solution with no recourse to programming languages. How to achieve maximum effectiveness of the model with minimal investment in its development
    • Ensembles: Development of a model for predicting property prices in the city of Oporto

Locations and Dates

  • Porto
    • Date: 07th and  08th of Februaryof 2022, from 09h00 to 18h00
    • Address: Rua Engenheiro Ferreira Dias, 161 - Porto
    • Note: This training is live-streamed
  • Lisboa
    • Date03th and  04th of March of 2022, from 09h00 to 18h00
    • Address:TBA
    • Note: This training is live-streamed

The Trainer

Nuno Nogueira is a business intelligence and analytics consultant in several companies in Portugal and abroad. He is the founder and CEO of Swell and the author of the book:  "Power BI para gestão e finanças" from FCA Publisher.