The Looker Query Insights Extension is a powerful custom visualization tool designed to summarize and present insights from any query created in a Looker explore. This extension provides a flexible and customizable interface, allowing users to tailor both the model parameters and UI to their specific needs. Inspired by this awesome Looker Hackathon Project, check it out!
- Custom visualization for summarized insights
- Compatibility with any Looker explore query
- Fully customizable model parameters and user interface
- Seamless integration with Looker, BigQuery, and Vertex AI
- Flexible frontend and backend architecture
only highlighting key files
📦 looker-query-insights/
├─ query-insights-backend/
│ ├─ README.md # instructions for Backend installation
│ └─ terraform/
│ ├─ variables.tf # GCP Project variables
│ └─ main.tf # GCP resource provisioning
└─ query-insights-extension/
├─ README.md # instructions for Extension installation
├─ manifest.lkml # the file required for configuring this extension in a Looker LookML project
├─ .env # where required environment variables are added
├─ package.json # where dependencies and startup/build commands are described
└─ src/
├─ index.js # where the Looker Extension SDK is first loaded and the app is mounted
├─ constants.js # where the Looker Viz Config Options are configured
└─ composables/
└─ useLookerVertexMessage.js #Handles Looker Query to Vertex for Prediction
Getting started involves (in order):
- Backend Setup. To note, if you've installed the Explore Assistant and have it configured in your Looker instance the Backend setup can be skipped. You can use the same backend resources for this extension
- Frontend Setup
This Looker Extension is mounted as a Custom Visualization, meaning that once installed you can access it from any Looker Explore as a visualization offering. Additionally it provides two tabs for editing the visualization as an end user: LLM configuration & UI Customization.
In the Explore Visualization config, the LLM tab will allow you to customize the prediction request with 3 parameters:
- Set Model Temperature (customize variability of reponse)
- Set Prompt Type (summary of results vs. forecast of time-series data)
- Set Query Description/Context (pass in any additional user provided context to the prompt sent to the llm)
In the Explore Visualization config, the Display tab will allow you to customize the look and feel of the visualization with 7 parameters: theme color, background color, card color, header color, shadown color, text color, and image url.
Inspired by this awesome Looker Hackathon Project by Juventin.
