
Quick introduction message to the course.
How to setup your environment for this course.
Explore data understanding through visualizations and distribution comparisons, showing how call duration, campaign contacts, and prior outcomes influence conversions.
Explore data preparation in CRISP-DM: clean and transform data, handle missing values, convert text to numbers, apply one-hot and CatBoost encodings, and prepare data for modeling.
Evaluate multiple models using F1 and accuracy in the crisp-dm framework, identifying catboost tuned as best without duration, while stressing variable impact and error analysis for improvements.
This lecture shows the preview of our final App that will be deployed.
Quick update on Experiments to show the autolog function.
In this lecture, learn how to register a model in MLFlow for deployment.
Create a streamlit app to deploy predictions from the best mlflow-registered model by loading the latest version, offering interactive inputs, and visualizing the prediction probabilities.
Learn how to add the application to a Docker container.
Learn how to push the Docker container created into the cloud.
In this hands-on course, you’ll learn how to execute a full data science project using the CRISP-DM framework, an industry-standard approach that guides you from understanding business needs to deploying your final model. Whether you're new to data science or seeking to expand your skill set, this course provides a practical, end-to-end experience that mirrors real-world project workflows.
Throughout this mini-course, we’ll cover each stage of CRISP-DM in detail, using Python to demonstrate essential techniques in data exploration, feature engineering, model training, and deployment. Starting with Business Understanding, you’ll learn to translate business challenges into actionable data science objectives. Then, we’ll dive into data preparation, exploring methods to clean and analyze data effectively, preparing it for modeling. You’ll work with real datasets and apply feature engineering techniques to make your model more accurate and insightful.
In the Modeling phase, we’ll select, train, and evaluate machine learning algorithms, optimizing them to create a robust solution. You’ll learn validation techniques to ensure your model’s performance and reliability, even in production environments. Finally, in the Deployment phase, we’ll cover how to prepare and deploy your model, so it’s ready for real-world use.
By the end of this course, you’ll have a solid foundation in CRISP-DM and the hands-on experience to confidently approach data science projects in a structured, methodical way. Join us to build real-world data science skills and make an impact with your analyses!