Back to All Events

How to Model Google Analytics Data with Dataform

How to Model Google Analytics Data with Dataform

About This Course

Are you ready to unleash the full power of BigQuery on your Google Analytics data? This course will show you how Dataform can be used to deploy a data model that transforms your raw Google Analytics data into a format that is ready for visualization and advanced analysis. At the end of this course you will have three things:

  1. A working data model that transforms your raw Google Analytics data incrementally as new tables arrive daily.

  2. All of the skills and resources needed to customize the model to your needs, including a series of Looker Studio dashboards that help you quickly start working with the data.

  3. Access to a community of previous course participants (in the TLC Slack) where you can collaborate with your peers as needed.

Who is it for?

This course is for the analyst who is ready to go beyond the reporting available in the GA4 user interface. You do not need to be a data engineer, but you do need to know enough SQL and Javascript to follow along (we will provide code for you to copy/paste). You will also need to understand the basics of Google Analytics 4 and how data is collected.

We strongly recommend completing Get Started with Google Analytics in BigQuery before taking this course.

Prerequisites

Before the course begins you will need to create a Google Cloud Platform project that is linked to a billing account (it is extremely unlikely that you will be charged any fees during this course, but Google requires a credit card just in case). If you need help with this process then you can follow the steps HERE.

We will start out by deploying our model to the sample Google Analytics dataset provided by Google, but if you have your own Google Analytics 4 property that is already linked with a Google Cloud Platform project we will help you deploy the model there as well.

Schedule

The full course includes 10.5 hours of content. Private teams may choose to complete this in 2 full days, but individuals taking the course through our Test & Learn Community learning groups will attend seven 1.5 hour sessions for 4 weeks.

  • Session 1) Introducing Dataform
    In session 1 we will show you how Dataform can be used to orchestrate your queries into a data model. We will cover the key concepts and make sure everyone has deployed a working model.

  • Session 2) Setting the cornerstone with the flat events table that is customized to your settings
    Once the basics are out of the way we will focus on converting the raw GA4 data into a flattened events table that is customized to your settings. Dataform will be configured to incrementally append new data to this table as it arrives daily.

  • Session 3) How to handle reporting identity
    With the flat events table in place we will discuss how you can create a new user identifier that recreates the selected reporting identity in Google Analytics. Then we will create a table that uses this identity to create user-scoped attributes.

  • Session 4) How to recreate the “Navigation Summary” report from UA for pathing analysis
    In session 4 we will show you how the work we did in the prior 3 sessions has enabled pathing analysis at the user and session levels. We will also create a Looker Studio dashboard that can be used to visualize your pathing analysis.

  • Session 5) Attribution modeling, and how to replicate the session attribution available in the user interface
    Session 5 introduces session-scoped attributes. We will replicate the session traffic attribution displayed in the “User Acquisition” and “Traffic Acquisition” reports and create a table that generates session-scoped metrics/dimensions (such as landing page).

  • Session 6) How to create segments and enable Ecommerce reporting
    Now that we have all of our user and session-scoped attributes, we will create a method for generating segments. We will also show how to create a separate table for Ecommerce reporting that handles item-scoped dimensions and metrics.

  • Session 7) How to Build a Semantic Layer
    In our final session we will wrap up any outstanding questions from prior sessions, and use Looker Studio to create a shareable dataset that can be used as a semantic layer.

What it isn’t

  • This is not a full deep dive into Google Cloud Platform and BigQuery. We will cover the specifics that you need to know to accomplish the goals of this course.

  • Not a SQL course, but you do not need to be an expert in SQL — If you can copy/paste then you’ll be able to follow along

  • No prior experience expected or necessary, but familiarity with Google Analytics concepts is assumed


How to Sign Up

Private Training for Teams
Click the button below to schedule a meeting with a member of our team so that we can discuss your needs.

Cohort Training for Individuals
Click the button below to see when the next course will begin through our partnership with the Test & Learn Community.


See Clips from this Course

Previous
Previous
January 1

Get Started with Google Analytics in BigQuery

Next
Next
January 2

Master BigQuery Identity Access Management