Here at DVELP we recently built a Looker Dashboard to monitor our progress towards our organisational goals. Here’s how we did it.
When you’re setting out to build a Looker Dashboard from scratch, step one is to decide what data you want to show and how to get it into Looker from where it lives.
We wanted to show data from GoogleSheets, Harvest, Hubspot and Trello to feed in our Dashboard. Connecting these services directly to Looker is tricky, but using a data warehouse with an integration tool allows you to get set up quickly.
Stitch has pre-written integrations with many services that make it easy to suck in data via the services’ APIs. We got Stitch to synchronise our data sources with our Amazon Redshift data warehouse. We then built visualisations of the information in our data warehouse. Every time you update your Dashboard, Looker queries that data warehouse. This is important, because if your synchronisations have not run in a while, you could be looking at outdated information. To make sure that doesn’t happen, stitch allows you to configure how often data is synchronised.
Navigating Looker revolves around the below banner that you see when you’re logged it. It is structured so that new users can ease their way in, working through the menus from left to right. If your organisation already has its data set up, this works well. However, if you’re setting up Looker with new data models, the order of operation is inverted.
Once your data feeds into Looker, there are five steps to get you from the raw data to your completed dashboard.
Configuring your View: here you tell Looker how to interpret your database. You configure which columns you want to use, what data types they are and how to aggregate numerical values when they’re grouped in a query
Configuring your Explore: here you tell Looker about the relationships that exist between tables in your database. This is done by defining how to join different tables together, much like in the SQL querying language
Querying your Explore: once you have your tables and relationships defined, you can use the Explore function to query your data. This is where the real power of Looker lives. Explores are like a GUI for writing queries that allows you to manipulate your data very quickly. This makes data analysis a breeze, because you don’t have to interrupt your analysis continuously to figure out how to write that complicated SQL query....
Customising your Look: once you have the data you need you can start customising your Look. This is basically a visualisation (read chart) of the data you’ve selected. Here you define chart types, axes labels, titles etc.
Building your Dashboard: once you have created and saved a few Looks, you’re ready to combine them into a Dashboard. Et voila! You’re all set.
As you will have noticed, there is a lot of (probably new) vocabulary to understand about Looker. Being clear on these definitions is essential - not least for describing your set up to Looker’s excellent customer service, which is available via Looker’s inbuilt chat function, super responsive and super helpful. We’ve summarised this vocabulary below. It is laid out according to where you would find it in the Looker banner. In orange you’ll see mnemonics that may help you remember their definition if you’re familiar with traditional SQL database architecture.
We hope you found this quick intro useful. If you have any trouble setting something similar up, please feel free to get in touch.