Hi, Evan here, and in this module, we'll explore in great detail, some of the cool things that you can do with SQL inside of BigQuery. Now I know what some of you may be thinking. SQL, isn't that just used for selecting and returning records for my database? Well, as the title shown here as already probably spoiled it, you can now build machine learning models with SQL too. It's literally my favorite feature in Google Cloud Platform, so I'll spend lots of time walking you through how you can create and analyze the performance of your ML models right within BigQuery where your data lives. But before we jump headlong into ML, we'll first introduce you to BigQuery as a service which allows you to have a petabyte-scale, analytics data warehouse. You'll soon learn that BigQuery is actually two services in one, a fast SQL Query Engine and fully managed data storage. Then, we'll show you some of the other cool built-in features, like using GIS functions on your Geographic data, and how you can even visualize those insights on maps. Next, we'll expand on our course theme of applying MLT or datasets, by looking at how to choose the right model type for your structured data. Lastly, we'll build a custom model using just SQL with BigQuery ML. I'll also show you how I commonly organize my ML projects and some advanced BigQuery ML features that let you do things like see which fields the model thinks are most important in making those predictions. Sound good? Let's get started. BigQuery is designed to be an easy-to-use data warehouse. We can focus on writing SQL statements on small or large datasets without worrying about infrastructure. If you've never written SQL before, I'll also provide resources and labs to get you up to speed as we go. So that's point Number 1, it's serverless. BigQuery's default pricing model is pay as you go. Where you pay for the number of bytes of data that your query processes and any other permanent data that's stored inside of BigQuery. Now there is some magic built-in like automatic caching of query results, so you don't end up paying for the same query returning the same data twice, which is cool. If you want to have a set bill every month instead, you can subscribe to flat tier pricing, where you get a special reserved amount of resources for your dedicated use. Data in BigQuery is encrypted at rest by default. You can also specify the geographic locality of your data if you need to meet things like regulatory requirements. Controlling access to your data can be as granular as specific columns, say any column tag with PII, Personally Identifiable Information or specific rows, like if your marketing team only needs access to see certain rows in one of your tables. BigQuery works in tandem with Cloud IAM to set these roles and permissions at a project level, and then inherited down to the BigQuery level. We'll discuss data access in detail a bit more later. SQL as a language has been around since the 1970s, and just watching the functionality added to the language over time has been awe-inspiring. You can now perform those GIS functions like distances from the lat long points and much more. You data itself, think of your datasets. It most likely has some kind of geographic component like city, state, zip code, latitude, longitude. So now it's high time to unlock those additional insights. Lastly, BigQuery as both a data warehouse and an Advanced Query Engine is foundational for your AI and ML workloads. It's common for data analysts, engineers, and data scientists to use BigQuery to store, transform, and then feed those large datasets directly into your ML models. This is a huge leap over training ML models on just a few small samples of your data locally on your laptop or desktop. You can now train on all the data that you have available. That's big news for ML. That's the elastic data warehouse nature of BigQuery for ML datasets. Beyond that, as you've seen in the demo, you can now write ML models directly in BigQuery using SQL. This is a great start for modeled prototyping, as you can quickly engineer what features to use right where your data lives. So what does a typical data warehouse solution architecture look like? Take a look at the green box. BigQuery is the analytics engine that sits at the end of a data pipeline like Cloud Dataflow. It stores all the incoming data from the left and allows you to do your analysis and your model-building. On the far right, you can see the myriad of visualization another analytical tools that you can connect through BigQuery as a backend data source. For ML engineers, once your dataset is in BigQuery, you can easily call it from your IPython ML notebooks in the cloud with just a few commands. If you're a business intelligence analyst, you can connect out to visualization tools like: Data Studio, Tableau, Looker, QlikView, and more. Lastly, and worth mentioning here, if you have a team of analysts who prefers to work in spreadsheets, you can now query your smaller huge BigQuery datasets directly, all of that data directly from Google Sheets and perform common operations like PivotTables, and more on the entirety of your dataset. So no more limitations of rows inside of sheets. So new feature, I'll send a link so you can check it out. But the key takeaway is that BigQuery is a common sink or staging area for your data analytics workloads. Once your data is there, your data analysts, business intelligence developers, and ML engineers can then be granted access to your data to start creating their very own insights.