Looker is a frontrunner among business intelligence tools, perfectly suited to exploring data of any size and performing on-demand analytics. If you haven’t used a tool of the same caliber, it can be a sharp departure from how you normally approach analytics, but for the better.
This article will be a comprehensive introduction to Looker, a discussion on how it can benefit your organization, a tutorial on how to leverage Looker’s modern analytics platform, and more.
Looker is a business intelligence and big data analytics solution that helps organizations pull useful insights from their data.
Users can choose between customer-hosted (self-service) and Looker-hosted deployments. The self-hosted option offers complete control over infrastructure administration, albeit at a higher initial cost and with hands-on maintenance. The Looker-hosted option is not a bespoke model but frees you from active management and updates.
Looker was acquired by Google in 2019 and became part of the Google Cloud Platform in 2020. Existing customers and partners at the time continued to support the tool, so the list of integrations only grew from there.
The pricing model is quote-based.
Looker, like all self-service BI tools, brings analytics closer to business users. Instead of requiring a complex IT environment, Looker sits on top of your SQL database and uses built-in connectors.
It usually takes developers/engineers to handle the initial integration (into existing business processes and tools in the Google ecosystem) and the modeling configuration. But then, non-technical users can pull the insights they need without assistance from the IT department. That ultimately creates a collaborative environment for different stakeholders, once impossible for day-to-day activities.
Looker breaks down barriers to insights through the following:
Looker Blocks are prebuilt data models (or pieces of code). By reusing the work of others, users can accelerate the development of analytics, insights, workflows, and applications.
To use a block, be it interactive visualizations or pre-modeled external data, go to Looker Marketplace and customize your chosen block to your specifications.
Looker Components are prebuilt pieces of user interface code. Not only do they simplify development, but they also minimize maintenance and overhead and enhance data experiences.
There are also components for rich filter functionality in apps and embedded dashboards: sliders, tag lists, radio buttons, etc.
Looker Alerts are tied to configurable actions. Users can specify the conditions for data and receive an alert if the conditions are met or exceeded. Looker will automatically check data at the required time frequencies.
The Looker Extension Framework is a fully hosted platform that takes care of certain aspects of web development, such as authentication, access control, API access, and more.
The goal is the same—simplification of processes. In this case, it’s building custom JavaScript data applications.
Looker’s native app is built for on-the-go access to data. It also allows users to seamlessly switch from desktop to mobile (with limited functionality) and share content.
Looker runs on Microsoft Azure, Amazon Web Services, and Google Cloud. With 60+ database dialects and 60+ SQL dialects, Looker also connects to your data stored in on-premises systems.
Instructions are available for 45+ SQL dialects, from Google BigQuery data warehouse to Oracle data lake architecture.
As mentioned, there are two architecture solutions: customer-hosted and self-service. There are different installation and deployment steps involved, so let’s go through them separately.
To install Looker for a customer-hosted deployment, you need to:
With this option, Looker provides an instance of the application in a shared virtual private cloud. As you’ll see, getting started with a looker-hosted instance is much easier:
Looker can be part of this popular modern data stack combination:
Of course, dbt with Snowflake is not the limit for Looker. You can browse applications, blocks, and custom plug-ins on the Marketplace Directory. Here are a couple of examples:
It’s important to note that some integrations are free and installed with a Git URL. But some are paid; for example, the Hubspot Segment integration costs $99/month for a once-a-day refresh.
Explores are starting points for all queries, each built to cover a particular subject area. For example, a business with an e-commerce model can select Explores like Orders, Order Items, Products, and Users to get data on and answer questions about each category.
Here is how you can query the Order Items Explore:
Some Explores have a Quick Start analysis option, which allows you to quickly populate fields in modeled queries. The prebuilt analysis options for an Order Items Explore are Order count by month, CA order count by month, and Order count by state by month.
Another key moment is source integration. You will likely want to populate your dashboards with data imported from different connected platforms. For this, you need to add resources to the tool:
User-defined dashboards are created and edited by most business users and Looker developers. The functionality is available in the Looker UI, but before you can begin, you must have the following permissions:
Keep in mind that you can no longer create legacy dashboards from scratch (in Looker 7.18 and up). Instructions below explain how to manage the new experience. But if you specifically need legacy dashboards, you can revert a regular dashboard to a legacy one (provided the Looker admin has enabled the feature).
This is the standard, most common way to create a dashboard (rather similar to creating a Looker report).
These simple actions will create a new blank dashboard, which you can customize as you want.
This process takes slightly longer because it has more inputs right away. The Look or Explore will be saved as a query tile on the dashboard.
Now, you can add tiles to the dashboard. The first one will take up the entire width of the dashboard, and all other ones will be automatically resized, but you can change the size, as well as edit the name, the data visualization within it, or the Looks. The rows of tiles are added as necessary.
There are three types of tiles:
To add a visualization, you need to:
After you create the tikes, you can still add filters for all tiles or select tiles. New filters will narrow down the results for data, but each filter must show at least one query tile or Look-linked tile; otherwise, it won’t be added to the dashboard.
In the blue toolbar, you can configure dashboard settings. For example, set the timezone for each tile. And you can change the dashboard description.
LookML dashboards are created by a select group of developers and are written and edited in a YAML-based dashboard file. These are the permissions required:
Also, any data used in the dashboard requires access to the LookML models for it.
Here is how you create a dashboard file with the extension .dashboard.lookml:
A new LookML dashboard file is already pre-populated with several basic dashboard (the layout, preferred viewers, tile sizes, etc. ) parameters and element parameters (the appearance and function of dashboard tiles, text, and buttons). But you can edit the file within the IDE, an integrated development environment for LookML developers, as needed.
You can also add visualizations by building queries on the Explore page and pasting the LookML under the elements parameter. To add filters, you can hard-code them into the dashboard elements or create the filters that users interact with and apply them using the listen element parameter.
A dashboard can’t serve its purpose unless it’s designed with effective performance in mind. The first thing you should focus on is building performant queries. Check out the backend tips for optimizing the underlying SQL query performance here.
As you continue building, refresh the page to confirm every element is working as intended and performance is not lagging. And when you are finished, test the performance and troubleshoot as needed. If you’re unsure what to fix or how to do it, Looker Support can lend a hand.
Businesses, from startups to enterprises, use Looker to uncover value in their data. What’s more, the level of customization and the amount of support make the solution suitable for practically any industry or company department.
The official website points out several industries with the highest Looker adoption:
Looker’s real-time insights can be useful in departments such as:
Before summarizing what we’ve learned about Looker, let’s list its strengths and weaknesses.
Pros | Cons |
|
|
Looker is one of the most robust and innovative pieces of software for data analytics. But it’s not perfect for every business user and has a lot of powerful competitors. Consider your business needs and whether Looker justifies its expense before committing to it.
Hopefully, this guide helps you understand the basics of Looker’s features, common uses, strengths and weaknesses, and the overall operations of Looker. If you want to integrate it into your business processes, you’ll definitely need to gain a more in-depth knowledge of the tool. Many users report a steep learning curve.
You need this guide, because it includes:
Turn it on to get exclusive guide on modern data stack
Emails suck. This newsletter doesn’t
Subscribe to the newsletter and get the most useful guide on modern data stack
The newsletter
You will also receive other useful materials on data analysis hacks with case examples from our company.