Quickstart: Conversational Analytics with GCP Billing and Looker

Key Concepts

• Looker Block: Pre-built LookML modules designed for rapid deployment of analytics solutions.
• Conversational Analytics: Interacting with data using natural language prompts.
• Persistent Derived Tables (PDTs): Pre-calculated data tables stored in the database to improve query performance.
• Looker with Gemini: Integration of Google’s Gemini AI model within Looker for enhanced conversational analytics.
• Role-Based Access Control (RBAC): Security practice of restricting system access to authorized users.

Streamlining Google Cloud Cost Management with Looker Blocks

This video demonstrates how to implement a streamlined solution for Google Cloud Platform (GCP) billing data analysis using a pre-built Looker Block, specifically designed for cloud cost management. The core problem addressed is the friction created by frequent ad hoc requests for billing data, and the security risks associated with granting broad data access. The solution focuses on empowering teams with self-service analytics through conversational interfaces, while maintaining secure, role-based access control.

Prerequisites & Setup

The initial phase involves preparing the GCP billing data and the Looker environment. Specifically, three key steps are required:

1. Data Export: Google Cloud billing data must be exported from GCP and stored within a database. The video specifically mentions a guide for BigQuery, indicating it’s a common target for this data.
2. Scratch Dataset: A dedicated dataset within the database needs to be created to house Persistent Derived Tables (PDTs). PDTs are pre-calculated tables that improve query performance by storing the results of complex calculations.
3. Connection Configuration: A Looker administrator must establish a connection between Looker and the database containing the exported billing data and the scratch dataset.

Installing the Cloud Cost Management Block

Once the prerequisites are met, the installation process is straightforward via the Looker Marketplace:

1. Access the Marketplace: Navigate to the “shop” icon (marketplace) in the top right corner of the Looker interface.
2. Search & Install: Search for “cloud cost management Google Cloud” and select the corresponding block. Click “install.”
3. Configuration Parameters: The installation requires four key parameters:
   • Connection: Specify the database connection previously configured by the administrator.
   • Recommendations Export Table: Enter the name of the table containing recommendations data (format: your_dataset_name.recommendations_export). This table is used for a single dashboard within the block.
   • Detailed Billing Table: Input the name of the table containing detailed billing information (likely containing “export resource” in the name). This table provides the most granular data.
   • Pricing Table: Provide the name of the table containing pricing data (format: your_dataset_name.cloud_pricing_export).
4. Installation Completion: After entering the parameters, click “install.” The video emphasizes that this process saves “weeks of SQL modeling.”

Utilizing Conversational Analytics

Following installation, the block unlocks conversational analytics capabilities through the “Conversations” feature in Looker.

1. Accessing Conversations: Open the Looker menu and select “Conversations.” Choose the “billing” explorer.
2. Enabling Features: If “Conversations” is not visible, the administrator needs to enable “Looker with Gemini” and configure appropriate permissions.
3. Role-Based Permissions: A default role, “conversational analytics user,” should be added to the relevant user group to grant access to the billing model data. This ensures secure, role-based access.
4. Interaction: Users can interact with the billing data using natural language prompts within the chat interface. Suggested prompts are provided to guide users. The system responds with interactive graphs and detailed answers.

Benefits and Impact

The implementation of this Looker Block offers several key benefits:

• Reduced Ad Hoc Requests: Empowers teams to self-serve their billing data needs, minimizing the number of ad hoc SQL requests directed to the analytics team.
• Enhanced Security: Eliminates the need to grant broad console permissions, maintaining secure access to sensitive billing data through role-based controls.
• Improved Data Understanding: Facilitates deeper data exploration and understanding through conversational interfaces and interactive visualizations.
• Increased Efficiency: Frees up the analytics team to focus on more complex analysis and insights, rather than fulfilling routine data requests.

Quote: “Now they can come to you with deeper understanding of the data and more informed analysis requests. This is how you can reclaim your time and empower your team.” – Speaker

Technical Vocabulary

• LookML: Looker’s modeling language used to define dimensions, measures, and relationships within the data.
• Explorer: A pre-defined starting point for data exploration within Looker.
• Persistent Derived Tables (PDTs): Pre-calculated tables stored in the database to improve query performance.
• Gemini: Google’s AI model integrated into Looker to power conversational analytics.

Conclusion

The video effectively demonstrates how a pre-built Looker Block can significantly streamline GCP billing data analysis. By leveraging conversational analytics and robust role-based access control, organizations can empower their teams with self-service data access, reduce friction, and improve overall efficiency. The key takeaway is that utilizing Looker Blocks can drastically reduce the time and effort required to build and maintain complex analytics solutions, allowing teams to focus on deriving valuable insights from their data.