If you are a beginner python programmer aiming for higher skills, stronger base knowledge, and a deeper understanding of the discipline of data science, and if you enjoy learning from books, here we have some suggestions for you.
Python for Data Analysis
Python for data analysis is written by the creator of the Pandas project, Wes McKinney. It is one of the best resources if you want an in-depth introduction to Pandas. Ask stated on the cover, the book is about data wrangling: how to clean, re-structure, store and tame raw data with python.
This one covers the same material as the python handbook above and even has been published in the same year, 2017. but it does not include a chapter on machine learning, instead, in 528 pages, it puts more focus on Pandas and Numpy libraries. The chapter structure is goal/use/purpose? Based instead of tool-based.
Python Data Science Handbook
Python Data Science Handbook was first written as a jupyter notebook, providing the ultimate experience of hands-on coding and learning for free here (https://jakevdp.github.io/PythonDataScienceHandbook/) on GitHub. It is 548 pages, although a little long, it covers nearly everything you need to learn as a beginner data scientist in a simple and practical manner. The five chapters discuss IPython, Numpy, Pandas, Matplotlib, Machine learning in bite-sized subchapters. To complete the set of your Python armor, the book also introduces you to ScikitLearn. If you’re a rookie, a bit familiar with basic python coding, and are now searching for ammunition to start data science, this book is suitable for you.
Data Science from Scratch by Joel Grus
Data Science from scratch teaches data science from scratch: no tools, no libraries, no mathematical prerequisites. Throughout this text, the reader gains insight into machine learning basic concepts and familiarizes with the use of math and programming in it, without getting stuck in tools and API without understanding what really is going on in the heart of a library call, or how machine learning really works.
This book starts with a self-supplementing crash course with python itself. This states how much prerequisite knowledge is required to understand it: nearly none. In later chapters, it introduces linear algebra, statistics and probability, known machine learning methods like k-nearest Neighbors, Naive Bayes, linear, multiple and logistic regression, clustering, neural networks, deep networks, network analysis, MapReduce, databases, and a chapter on data ethics! All in just 330 pages.
This book doesn’t teach math. If you’re looking for extensive theoretical knowledge, you must know that nearly half of Data Science from Scratch is code. It also doesn’t do libraries (except some Matplotlib for visualization), so if you’re looking for a guide to Numpy or Pandas, you’re at the wrong place.
Think Stats by Allen B. Downy
From the author of the well-known Think Python, this book is less tool-based and mainly emphasizes the underlying concepts of known ML methods, teaching the reader how to think and solve problems like a data scientist. It’s short, concise, and contains an extensive set of exercises.
A full practical case study over a real-world dataset is used to demonstrate concepts throughout the book. Readin Think Stats helps you find a deeper, more practical understanding of probability and statistics by trying computational simulation experimenting with data generation and distributions. The book enables you to use any interesting dataset that is accessible, not just preprocessed and cleaned datasets- for your analysis projects.
The author is an experienced college teacher of software and programming courses, and the material of Think Stats is refined by multiple semesters of being thought in classrooms. In addition to statistics and python, it’s going to give you a good problem-solving sense. We especially recommend it to undergraduate students seeking to start learning data science.
Hands-on Machine Learning with Scikit-Learn, Keras & TensorFlow
Today programmers can implement machine learning methods into their applications even without knowing anything about the underlying concepts. That’s not the level of knowledge Aurelien Geron intended to supply the reader with, but he has made minimal theory and mathematics an important feature of his work.
The book enables you to apply Sckikitlearn and TensorFlow, Keras pandas frameworks, for problems and purposes ranging From data cleaning to simple linear regression to deep neural networks, introduction to supervised and unsupervised learning. Each chapter includes programming exercises and all the answers and codes are available on Github. You implement an End to End project throughout the book, going through all stages of a data science project from modeling to deployment to google cloud.
Reading Hands-on Machine learning with Sckikt-Learn, Keras & TensorFlow gives you a problem-solving mindset and familiarity with a set of powerful python tools, all put in order to make you able to approach data science problem with keen vision and full hands