Core Concepts & Mental Models

What are the concepts behind Toucan.AI and how can they be leveraged to build analytics features?

TL;DR

Toucan AI gives you a toolkit of concepts and components to help your team build, manage, and embed analytics—whether for your product or your internal workflows.


Core Building Blocks

Organizations and Workspaces

  • Organization: A secure group of users with shared access to databases, charts, dashboards, and permissions.

  • Workspace: The collaborative environment within an organization where analytics content is created and managed.

Roles & Permissions

Toucan AI uses role-based access control to keep your data secure and your workflows organized:

  • Admin: Full access to all features, including user and data management.

  • Maker: Can create and modify charts, dashboards, databases, and tables, but cannot manage users or organization settings.

  • Explorer: Read-only access to shared content; can view dashboards and interact with filters, but cannot create or modify content unless granted extra permissions.

Data Sources and Datasets

  • Database: Where your data lives—Toucan AI supports SQL databases and data warehouses.

  • Table: Structured data within a database (rows = records, columns = attributes).

  • Row-Level Security: Attribute-based rules that restrict data access at the row level, automatically enforced in every query.

Semantic Layer and Metrics

  • Semantic Layer: Define business concepts—metrics, dimensions, and relationships—on top of raw data for consistent, reliable analytics.

  • Metrics: Quantitative values (like revenue, user count).

  • Dimensions: Qualitative attributes (like region, product) for grouping or filtering metrics.

Charts, Dashboards, and Layouts

  • Chart: A visual representation of your data (bar, line, donut, etc.), built from datasets and semantic definitions.

  • Dashboard: A collection of charts and filters, organized to answer specific business questions—customizable and shareable.

  • Filter: Refines dashboard data by applying criteria to one or more charts, based on unique values from selected columns.

AI-Assisted vs Manual Creation

  • AI-Assisted: Use natural language to generate charts, dashboards, and queries. Toucan AI’s AI suggests visualizations and insights based on your data and semantic layer.

  • Manual Creation: Build and customize charts and dashboards directly, with full control over data, metrics, and layout.

Embedded vs Internal Analytics

  • Embedded Analytics: Integrate dashboards and analytics features into your product, so users can access insights right inside your application.

  • Internal Analytics: Use Toucan AI’s interface for your own teams to explore and analyze data.

Authentication and Tokens

  • Token: Used for authentication when embedding Toucan AI visualizations. Generated via API with an API key, and can include attributes for row-level security and object access control.


Summary

These concepts are the foundation of Toucan AI. They enable you to connect data, model business logic, create visualizations, and deliver analytics—embedded in your product or for internal use—efficiently and securely.

Last updated