Feature engineering with Python
Find out what you will learn throughout the course (if the video does not show, try allowing cookies in your browser).
What you'll learn
► Multiple methods for missing data imputation.
► Strategies to transform categorical variables into numbers.
► How to handle infrequent categories.
► Variance stabilizing transformations.
► Multiple discretization techniques.
► How and when to handle outliers.
► How to create features from dates and times.
► Apply transformations with Python open source libraries.
☻ More than 17k students enrolled.
☻ More than 2.5k student reviews.
☻ Average course rating: 4.7 out of 5.
What you'll get
► 10+ hs. of video lectures
► Presentations, quizzes and assignments.
► Jupyter notebooks with code.
► Instructor support through Q&A.
► Access in PC and mobile.
► Lifetime access to content.
► Certificate of completion.
30 days money back guarantee
So you can buy with confidence.
Course description
Welcome to the most comprehensive course on feature engineering for machine learning available online.
What is feature engineering?
Feature engineering is the process of using domain knowledge and statistical methods to create features that make machine learning algorithms work effectively.
Feature engineering is key in applied machine learning. Raw data is almost never suitable to train machine learning models. In fact, data scientists devote a lot of effort to data analysis, data preprocessing, and feature extraction, to create better features to train predictive models.
What will you learn in this online course?
In this course, you will learn about missing data imputation, encoding of categorical features, numerical variable transformation, discretization, and how to create new features from your dataset.
Specifically, you will learn:
- How to impute missing values
- How to encode categorical features
- How to transform and scale numerical variables
- How to perform discretization
- How to remove outliers
- How to perform feature extraction from date and time
- How to create new features from existing ones
While most online courses will teach you the very basics of feature engineering, like imputing variables with the mean or transforming categorical features using one hot encoding, this course will teach you all of that, and much, much more.
You will first learn the most popular techniques for variable engineering, like mean and median imputation, one-hot encoding, transformation with logarithm, and discretization. Then, you will discover more advanced methods that capture information while encoding or transforming your variables, to obtain better features and improve the performance of regression and classification models.
You will learn methods described in scientific articles, used in data science competitions like those hosted by Kaggle and the KDD, and that are commonly utilized in organizations. And what’s more, they can be easily implemented by utilizing Python's open-source libraries.
Feature engineering with Python
Throughout the course, we will use Python as the main language. We will compare the implementation of feature engineering with the open-source libraries Pandas, Scikit-learn, Category Encoders and Feature-engine.
Throughout the tutorials, you’ll find detailed explanations of each technique and a discussion about their advantages, limitations, and underlying assumptions, followed by the best programming practices to implement them in Python.
By the end of the course, you will be able to decide which feature engineering technique you need based on the variable characteristics and the models you wish to train. And you will also be well placed to test various transformation methods and let your models decide which ones work best.
Who is this course for?
This course is for data scientists and software engineers who want to improve their skills and advance their careers.
Course prerequisites
To make the most out of this course, learners need to have basic knowledge of machine learning and familiarity with the most common predictive models, like linear and logistic regression, decision trees, and random forests.
To wrap-up
This comprehensive feature engineering course contains over 100 lectures spread across approximately 10 hours of video, and all topics include hands-on Python code examples in Jupyter notebooks that you can use for reference, practice, and reuse in your own projects.
Soledad Galli, PhD
Instructor
Sole is a lead data scientist, instructor and developer of open source software. She created and maintains the Python library for feature engineering Feature-engine, which allows us to impute data, encode categorical variables, transform, create and select features. Sole is also the author of the book "Python Feature engineering Cookbook" by Packt editorial.
Course Curriculum
- Variable characteristics (2:44)
- Missing data (6:46)
- Cardinality - categorical variables (5:04)
- Rare labels - categorical variables (4:54)
- Linear models assumptions (9:13)
- Linear model assumptions - additional reading resources (optional)
- Variable distribution (5:08)
- Outliers (8:27)
- Variable magnitude (3:09)
- Variable characteristics and machine learning models
- Additional reading resources
- Introduction to missing data imputation (3:51)
- Complete Case Analysis (6:46)
- Mean or median imputation (7:53)
- Arbitrary value imputation (6:39)
- End of distribution imputation (4:53)
- Frequent category imputation (6:56)
- Missing category imputation (4:05)
- Random sample imputation (14:14)
- Adding a missing indicator (5:25)
- Imputation with Scikit-learn (3:43)
- Mean or median imputation with Scikit-learn (5:27)
- Arbitrary value imputation with Scikit-learn (5:04)
- Frequent category imputation with Scikit-learn (5:09)
- Missing category imputation with Scikit-learn (2:50)
- Adding a missing indicator with Scikit-learn (4:06)
- Automatic determination of imputation method with Sklearn (8:24)
- Introduction to Feature-engine (6:25)
- Mean or median imputation with Feature-engine (4:51)
- Arbitrary value imputation with Feature-engine (3:30)
- End of distribution imputation with Feature-engine (4:46)
- Frequent category imputation with Feature-engine (1:38)
- Missing category imputation with Feature-engine (3:04)
- Random sample imputation with Feature-engine (2:00)
- Adding a missing indicator with Feature-engine (4:06)
- CCA with Feature-engine (6:47)
- Overview of missing value imputation methods
- Conclusion: when to use each missing data imputation method
- Categorical encoding | Introduction (6:49)
- One hot encoding (6:09)
- Important: Feature-engine version 1.0.0
- One-hot-encoding: Demo (14:12)
- One hot encoding of top categories (3:06)
- One hot encoding of top categories | Demo (8:35)
- Ordinal encoding | Label encoding (1:50)
- Ordinal encoding | Demo (8:08)
- Count or frequency encoding (3:11)
- Count encoding | Demo (4:33)
- Target guided ordinal encoding (2:41)
- Target guided ordinal encoding | Demo (8:30)
- Mean encoding (2:16)
- Mean encoding | Demo (5:31)
- Probability ratio encoding (6:13)
- Weight of evidence (WoE) (4:36)
- Weight of Evidence | Demo (12:38)
- Comparison of categorical variable encoding (10:36)
- Rare label encoding (4:31)
- Rare label encoding | Demo (10:25)
- Binary encoding and feature hashing (6:12)
- Additional reading resources
- Discretisation | Introduction (3:01)
- Equal-width discretisation (4:06)
- Important: Feature-engine v 1.0.0
- Equal-width discretisation | Demo (11:18)
- Equal-frequency discretisation (4:13)
- Equal-frequency discretisation | Demo (7:16)
- K-means discretisation (4:13)
- K-means discretisation| Demo (2:43)
- Discretisation plus categorical encoding (2:54)
- Discretisation plus encoding | Demo (5:45)
- Discretisation with classification trees (5:05)
- Discretisation with decision trees using Scikit-learn (11:55)
- Discretisation with decision trees using Feature-engine (3:48)
- Domain knowledge discretisation (3:52)
- Additional reading resources
- Feature scaling | Introduction (3:44)
- Standardisation (5:31)
- Standardisation | Demo (4:39)
- Mean normalisation (4:02)
- Mean normalisation | Demo (5:21)
- Scaling to minimum and maximum values (3:24)
- MinMaxScaling | Demo (3:01)
- Maximum absolute scaling (3:01)
- MaxAbsScaling | Demo (3:45)
- Scaling to median and quantiles (2:46)
- Robust Scaling | Demo (2:04)
- Scaling to vector unit length (5:51)
- Scaling to vector unit length | Demo (5:18)
- Additional reading resources
Frequently Asked Questions
When does the course begin and end?
You can start taking the course from the moment you enroll. The course is self-paced, so you can watch the tutorials and apply what you learn whenever you find it most convenient.
For how long can I access the course?
The courses have lifetime access. This means that once you enroll, you will have unlimited access to the course for as long as you like.
What if I don't like the course?
There is a 30-day money back guarantee. If you don't find the course useful, contact us within the first 30 days of purchase and you will get a full refund.
Will I get a certificate?
Yes, you'll get a certificate of completion after completing all lectures, quizzes and assignments.