How to Teach Looker - Intro to Looker, Using the Explore Interface, and Sharing Content

This article is written with the Looker implementation leader for your company in mind. However, it might be a useful reference for any Looker user learning how to use dashboards and explores in Looker.

An important part of implementing Looker is achieving buy-in from end users (sometimes called business users) at your company. These are the users who will be consuming dashboards and exploring content, but not necessarily altering LookML just yet. This article should prepare you to stimulate usage of both dashboards and explores by these end users. This is sometimes called a “Train the Trainer” exercise, meaning that once you understand this content you are ready to teach Looker exploration across an organization.

Introduction to Looker: The Joys of Browsing Saved Content

At the beginning of your training, it’s important to get the crowd used to getting around in Looker.

First things first, make sure everyone is actually logged into Looker. The training won’t matter if they haven’t sorted out access!

What is Looker?

Looker is a way for you to access your company’s data via a web browser. Simple! The navigation bar includes Browse and Explore, both of which we will cover.

Also point out the Help button on the navigation bar, to give users a place to go when you aren’t available to answer questions.

The User button is where they can change their user profile.

Browsing Spaces

A good place to start an overview of Looker is found by following the “Browse” link at the top. This will view the “Spaces” organizing saved content in your Looker instance. This area works just like any other folder-file structure, and analogies to Google Drive or Windows Explorer / Mac Finder are helpful. “Spaces are where you can access the saved content in Looker, whether saved by your analyst team or by other users. Saved content includes Looks and Dashboards.”

Looker Dashboards

Have the group open a dashboard in a shared space. Choose something useful, which will demonstrate interesting data and a few key dashboard patterns.

What is a Dashboard?

A dashboard is a collection of reports. Each one provides insight into how your company is operating, and each is based on the data your company generates or gathers. By grouping several reports together, a dashboard is a powerful place to inform decisions. Sometimes this sort of insight is called Business Intelligence. Your Looker instance contains many dashboards.

Viewing a Dashboard

Once you’ve navigated to a dashboard, how does it work? The first thing to do is make sure it has been run. There is a purple “Run” button in the top right, click it! After all the tiles have finished loading, information on how fresh the data is will be displayed next to the run button.

You can scroll through a dashboard, as well as within each tile individually (if the tile has more data than is shown right away).

Interacting with Dashboard Visualizations

There are several things you can do with the tiles in a dashboard.

  • Hover over a line, bar, or other visualization element. This will show a tooltip with things like the precise number, and any drill options (more on that below)
  • Clicking on legend elements will hide or show that part of the visualization
  • Clicking on table headers will sort by that column. Hold down the shift key to select multiple columns.
  • Sometimes maps can be zoomed in/out or panned, depending on how your analyst team set them up

Drilling

A special way to interact with reports in a dashboard is drilling. This action uncovers the data behind visualized numbers. For example, drilling into a count of orders will lead you to a list of those orders.

Any value in a table which has an underline (like a hyperlink) upon hovering is drillable. Some visualization elements (e.g. bars, lines, etc.) are also drillable, click them to find out! Your analyst team sets up which numbers have drill paths as well as what fields are displayed upon drilling, so contact them if you need this customized. These can even link to other dashboards or URLs outside of Looker! Way cool.

####Dashboard Filters
Dashboard filters are another special way to interact with a dashboard. If your dashboard has them (and not all do), they will be located at the very top. Any filters set here will impact multiple tiles in the dashboard. Which tiles the filter impacts is set by whomever created the dashboard.

The Gear Icon

19fbf22cc5d0941beb7e934987f7f48c20297c8c.png

The gear icon is found all over Looker. In the dashboard interface, there is a gear icon at the top which applies to the dashboard as a whole, as well as a gear icon for each tile. They do different things!

At the dashboard level, the gear is your portal to…

  • Download the dashboard as a PDF
  • Schedule the dashboard to be sent out at regular intervals (more on scheduling at the bottom of this article)
  • Clear Cache and refresh the data displayed in the dashboard

Each tile also has a gear icon, visible upon hovering over the tile. It can be used to…

  • Download the data displayed in the tile in a variety of file formats
  • View the original look which built that tile. More on Looks next!

Looks

So far we have spoken about the reports in a dashboard as “tiles.” You can also view individual reports by themselves. These are called Looks. Looks are the best way to make new content before organizing them into dashboards.

To view a Look, click on a gear icon at the tile level. Then, choose “view original Look.” You will be taken to the Look interface for that Look. You may also recall seeing a list of Looks in the Spaces interface at the beginning of the training. Looks are also stored as their own independent content, and can be accessed directly from the Space to which they are saved.

Looks have several elements:

  • Filters. They can be filtered using the filters at the top.
  • Visualization Settings. You can change the visualization type and settings by clicking the gear icon for the visualization.
  • Data. You can view the underlying data by clicking the “Data” header and expanding the data set which underlies the visualization.
  • Meta data about the Look. This information is on the right, and includes a description, a list of schedules associated with the look, and a list of which dashboards it is displayed in.
  • Gear icon. This has options like scheduling (more on that later), download data, and others.

There is one more link, the “Explore from Here” link at the top right of the visualization. This is the best part. For more information on how Exploring can be used to change existing Looks or create new ones, read on!

Practice

Notice several places to stop and have the new user actually perform some actions with Looker. This will help the training stick. Customize this experience for each user or group of users by demonstrating with content relevant to them.

Exploring: Finding Answers to Your Own Questions

Now that your audience is oriented in Looker, your task is to get them to move past just consuming dashboards and into using the explore interface. This will allow them to serve themselves with your company’s data unleashing the real power of Looker for end-users. The explore interface is a very powerful tool for asking additional questions beyond what is already saved in a dashboard or Look.

A shortcut for SQL-Savvy Users

Something to consider before getting started is your audience’s SQL knowledge as well as experience with Looker. After all, Looker is a modeling layer designed to convert user interaction directly into SQL, so there are clear connections.

Most end users do not know SQL and that is okay, they don’t need to learn SQL—Looker is designed for them. But, if they do, you can accelerate the skills transfer by invoking analogies to SQL. These are the relations between Explore elements and the SQL Looker writes:

  • Dimensions: included in the SELECT clause and GROUP BY clause
  • Measures: included in the SELECT clause using aggregate functions
  • Filtering on Dimensions: included in the WHERE clause
  • Filtering on Measures: included in the HAVING clause
  • Pivots: not an important part of the SQL query, only changes how the results are displayed.

Key Concepts for Exploring

Even without any SQL knowledge, end users can become proficient with interacting with the explore interface. The following concepts are key for their success.

Dimensions

A dimension is what you want information about. Tips that you are talking about a dimension are saying you want something “by x dimension” or for “each x dimension.”

Most critical is getting users to understand the “Law of Dimensions:" Choosing a dimension will return all the unique results for that dimension. Looker will then create one row in the result set for each unique record.

When multiple dimensions are chosen, Looker will return the unique combinations of these dimensions, NOT the unique results of each individually. One way to explain this is that Looker concatenates the dimension columns selected and groups by the new unique concatenated phrases.

Demonstrating with a data set small enough to fit on one screen helps learners visualize and retain this concept. A small excel spreadsheet or simple table of data can be helpful, as each column in the sheet or table represents different dimensions of the data. Once the concept is understood, it is not so difficult to apply to a larger, more realistic dataset.

Measures

If a dimension is what you want to know about, the measure is the “what” you are looking to learn. In plain language you would want to know “how many y measure for each x dimension.”

A good place to start when demonstrating measures is with a single measure selected and no dimensions. This will demonstrate that measures are typically mathematical aggregations of data such as sum, average, or max/min without complicating the explanation at first. The key here is to understand that measures aggregate or summarize multiple rows in the data into just one value.

Once combined with one or more dimensions, then this concept of “aggregating” or “compressing” multiple rows into one row helps to underscore the function of dimensions. The measure aggregation is taking the multiple rows which match one unique dimension result and summarizing them into one row using some aggregate function.

Filtering on Dimensions

A dimension filter removes data from the raw underlying table before any dimension-ing or measure-ing takes place. These filters are applied on a row-by-row basis by simply checking whether that row’s value for the dimension meets or does not meet the filter criteria. This concept is straightforward (once dimensions are measures are well understood) and usually provides few problems with learners.

Filtering on Measures

This concept is trickier than dimension filtering. Expect your end users to have some trouble here, especially if there are closely named dimension filters to confuse with the measure filters. The issue becomes how to choose when to filter on a measure instead of a dimension. A way to gain confidence here is to understand how a measure filter works. Then it should be easier to know when to use one in place of a dimension filter.

So, how do measure filters work? In order to filter on a measure, something first has to be done to the raw data before the filter can be applied. This is a key difference from dimension filters, which are applied right away. To explain this, point out that when applying filter criteria to a measure there is simply nowhere to find that information in the underlying data. For example, we have a table of orders in 2016 and want to see a count of orders for each month. If we want to only see months with more than 10,000 sales, that information is not in the underlying data. Looker must first perform the counting aggregation before any filtering on that measure can take place.

In short, use a measure filter if the action desired requires some aggregation first.

Pivots

Pivots are usually grasped easily by end users. Comparing them to Pivot Tables in excel can help, but even if they are not Excel users pivots are not a difficult concept. Pivots work in the same way as dimensions. Since pivots are actually pivoted dimensions, they adhere to the Law of Dimensions. The result set will be grouped by the unique combinations of all dimensions and pivoted dimensions. The only difference between choosing to pivot on a dimension rather than just adding it as a regular dimension field is the way the data will be visually rendered. Note that visualizations are impacted by whether a dimension is or is not pivoted.

Practice

This is super important. Don’t be afraid to move at a slow pace through practice questions.

Within reason, wait until all users have achieved the intended result before moving on. This enables all learners to benefit from inevitable nuances which are brought up during troubleshooting the issues encountered by slower folks.

In particular, focus on fields which look similar but one is a measure and one is a dimension, both in selecting and filtering since these test understanding most deeply.

Sharing: Presenting Explorations to Others

Once users are comfortable exploring, they will want to share the things they learn with other people at your company. This section walks through the practicalities of getting others to view their new content, as well as how to apply visualizations and custom table calculations.

Saving Looks, Sharing Links, and Creating Dashboards

A great exercise to do during a business user training is to have your audience do some exploring, and then save it as a look. Then have them practice sharing their Look via saving to a space or sending a link, adding it to a new dashboard, and even scheduling an email of the look or dashboard.

Dashboard filters are complicated, so this might be too advanced for your business users. However, the feature is very valuable to certain users, so gauge your audience on this one. If you do decide to teach dashboard filters be sure to take plenty of time to do so, since these are the closest that end-users will come to using variables in Looker.

Creating Graphs and Other Visualizations

Visualizations can be difficult to teach. An exhaustive dive into all the visualization parameters is impractical and would probably be pretty boring. In practice it is often easier to find a visualization example similar to what is desired, then use it as a reference while building the same visualization for another data set. There are several great examples in the learn.looker.com instance (request access to Learn from your Customer Success Manager or Jumpstart Consultant), or use examples in your own instance. The key here is to make sure your users have access to example visualizations to learn what is possible and how to investigate to see how it was done.

Adding Custom Fields as Table Calculations

In addition to teaching table calculation syntax, there are a couple important things to get across as well.

First, the table calculation only lives in that one Look. If an end-user finds themselves using a particular table calc in more than one Look, they should request their analyst (you!) to be added to the model.

Second, Table calculations differ from normal Looker behavior in that the data available to them is held in the Looker result set, not in the underlying database. This is different from normal exploration where the available data is everything included in the database (or at least the part of the database included in that explore). A good way to explain this is with an analogy of copy/pasting the result set into Excel, then manipulating it there.

A helpful demonstration of how table calcs only use the data in the result set is by creating a percent of total calculation.

  1. Pull a dataset which includes at least one dimension and one measure.
  2. Enter table calcs and add one called “Total measure_name” as sum(${measure_name}).
  3. Save and check out the table calc. Point out that it is the same on every row, and only added up the values in the result set.
  4. Enter table calcs again and create a second table calc called “Percent of Total measure_name” as ${measure_name}/${Table calc from step #2}. Remember to choose the appropriate value format.
  5. Voila! Save and check out the data again, and ask whether there are questions.
16 2 2,035
2 REPLIES 2

Hello,
This information is impressive; I am inspired with your post writing style & how continuously you describe this topic. After reading your post, thanks for taking the time to discuss this, I feel happy about it and I love learning more about this topic…

JamesGray
Participant I

very helpful! Thank you!

Top Labels in this Space
Top Solution Authors