Google Bigquery Ml Machine Learning In Sql (Without Python)

On Linear Regression example

What you'll learn
Create Machine Learning model and make prediction using only SQL code
Evaluate and interpret model prediction quality
Do Feature Engineering on different data types
Clean up and limit data source with understanding of consequence of it
Requirements
Basic knowledge of SQL
Description
The goal of this course is to learn how to create and use Machine Learning models right from the level of SQL query in Google BigQuery interface. You will also learn how to prepare data, how to interpret model results and how to make nice predictions using just one SELECT statement. You will work on a real data set - car sale offers in the USA, and the goal will be to predict the price of a car.The course consists of 7 sections and one bonus section. At the very beginning we will create an environment to work in. Next it would be good to see a little theory. Then we will straight jump into the first model creation. In further lessons we will try to improve our model performance by some hacks and tricks. This is essential for the course and we put the biggest pressure on that part. In the meantime you will get all needed resources and you will be able to practice all steps by yourself on your own free BigQuery account.In this course you will be working on your own end project. During the course, we will guide you on how to make every step of your own end project. After each practical lesson, you will have a homework assignment that will contribute to your big project. The project’s goal is to predict used car prices. Additionally, to motivate you to work and check if you have done your homework correctly, you will get a question in the quiz. By carrying out practical tasks, you will easily find answers.We’ve added a few lesson resources. Google glossary ebook that explains all basic definitions of a wide spectrum of Machine Learning. Please read them to systematize your knowledge. Other resources are cheat sheets which present a summary for each topic. It's a really nice source of condensed knowledge. Please use them to quickly look if you forgot some stuff. For practice lessons we add our SQL in resources. You can easily copy-paste and manipulate the code by yourself.Let’s get started with our journey of Machine Learning in SQL!

Overview

Section 1: Before start the Course

Lecture 1 Lesson 0.1 Course Introduction

Lecture 2 Lesson 0.2 First Thing To Do

Lecture 3 Lesson 0.3 Setting up BigQuery Sandbox

Section 2: Introduction - basic concepts and theory

Lecture 4 Lesson 1.1 What is Machine Learning?

Lecture 5 Lesson 1.2 What is Linear Regression?

Lecture 6 Lesson 1.3 What is Google Cloud Platform and BigQuery?

Lecture 7 Lesson 1.4 What is BigQuery ML?

Lecture 8 Lesson 1.5 BigQuery Data types

Lecture 9 Lesson 1.6 BigQuery SQL Fundamentals

Section 3: Creating first model and prediction

Lecture 10 Lesson 2.0 Section introduction

Lecture 11 Lesson 2.1 Business goal and model limitation

Lecture 12 Lesson 2.2 Data source description

Lecture 13 Lesson 2.3 BigQuery User Interface

Lecture 14 Lesson 2.4 Import data to BigQuery

Lecture 15 Lesson 2.5 Create model

Lecture 16 Lesson 2.6 Predict data

Lecture 17 Lesson 2.7 Model evaluation

Section 4: Data cleaning

Lecture 18 Lesson 3.0 Section Introduction

Lecture 19 Lesson 3.1 Removing useless columns

Lecture 20 Lesson 3.2 Data visualization with Google Data Studio

Lecture 21 Lesson 3.3 Histogram

Lecture 22 Lesson 3.4 Checking duplicates

Lecture 23 Lesson 3.5 Removing null values

Section 5: Feature engineering

Lecture 24 Lesson 4.0 Section introduction

Lecture 25 Lesson 4.1 Create new feature - car age

Lecture 26 Lesson 4.2 Create new feature - VIN number

Lecture 27 Lesson 4.3 Create new feature - Condition field

Lecture 28 Lesson 4.4 Create new feature - Model field

Lecture 29 Lesson 4.5 Create new feature - Geography

Section 6: Feature engineering - built-in function

Lecture 30 Lesson 5.0 Section introduction

Lecture 31 Lesson 5.1 ML.MIN_MAX_SCALER function

Lecture 32 Lesson 5.2 ML.FEATURE_CROSS function

Lecture 33 Lesson 5.3 ML.POLYNOMIAL_EXPAND function

Lecture 34 Lesson 5.4 ML.QUANTILE_BUCKETIZE function

Lecture 35 Lesson 5.5 ML.BUCKETIZE function

Lecture 36 Lesson 5.6 ML.NGRAMS function

Lecture 37 Lesson 5.7 Removing unimportant columns

Section 7: Hyperparameters tuning

Lecture 38 Lesson 6.0 Section introduction

Lecture 39 Lesson 6.1 L1 & L2 regularization

Lecture 40 Lesson 6.2 Automatic vs manual tuning

Section 8: Final prediction and model testing

Lecture 41 Lesson 7.0 Section introduction

Lecture 42 Lesson 7.1 Negative price

Lecture 43 Lesson 7.2 Using logarithm function

Lecture 44 Lesson 7.3 Test model quality

Lecture 45 Lesson 7.4 Final lesson

Section 9: Extra Section: Boosted Tree Algorithm

Lecture 46 Extra Lesson 0: Introduction

Lecture 47 Extra Lesson 1: Boosted Tree short theory

Lecture 48 Extra Lesson 2: Model train and predict

Beginner Data Analysts or students who want to start with Machine Learning using just SQL