What this book covers
Chapter 1, Understanding the AI/ML Landscape, provides an overview of the current state of data science as well as what tools you'll need to succeed.
Chapter 2, Analyzing Open Source Software, delves into the role of OSS in data science and how to decide what new OSS tool to use. You'll get a systematic checklist to look for in the next tool you evaluate.
Chapter 3, Using Anaconda Distribution to Manage Packages, covers how to manage packages with conda and Navigator. This includes how to create environments and create channels.
Chapter 4, Working with Jupyter Notebooks and NumPy, covers how to successfully turn notebooks into your daily driver to create data science value. We'll also go deeper into the powerful NumPy library to vastly speed up our operations.
Chapter 5, Cleaning and Visualizing Data, looks at the core techniques you'll need to shape data coming in to prepare it for model training. We'll cover areas such as imputing and also how we can visualize our data to gain a greater understanding.
Chapter 6, Overcoming Bias in AI/ML, looks at the many ways that naive ignorance can be present in our data and what we can do to avoid or correct these issues. You'll see what the real-world impacts are of a biased AI model.
Chapter 7, Choosing the Best AI Algorithm, goes into some of the major problem families that AI/ML models can help with, including regression and anomaly detection. We'll check out the algorithms you can use as well as the comparative rating for each.
Chapter 8, Dealing with Common Data Problems, looks at how you can identify and correct errors in your datasets, such as incorrect data entries. You'll also see how to scale your data and encode categorical features.
Chapter 9, Building a Regression Model with scikit-learn, walks you through a complete flow of building a regression model and how you can evaluate the results.
Chapter 10, Explainable AI – Using LIME and SHAP, goes further into the results of a model to be able to interpret and also explain how a model arrived at the results it did. Models that are interpretable by design and black-box models are covered.
Chapter 11, Tuning Hyperparameters with scikit-learn Pipelines, takes a more holistic approach and shows you how to leverage pipelines to create a flexible and repeatable process for data preparation and model creation. We'll cover how to use these tools to tune your hyperparameters to create a better model.