Skip to main content

Analyses

Technical Database Schema

For the transactional database fields and schema structures, refer to the app.analyses Table Documentation and the app.analysis_logs Table Documentation.

In Ravioli, Analyses are interactive projects designed to explore and query datasets. They represent the primary Contributor activity, enabling users to transform raw data assets into actionable knowledge.

Triggering an Analysis​

An analysis project can be initiated in a few ways:

  • Data Ingestion: Directly uploading a new dataset (e.g., CSV, Parquet) or connecting to a database.
  • Workspace Navigation: Selecting an existing data asset and initiating an exploration session.
  • Knowledge Context: Attaching Knowledge Pages to provide domain-specific metadata and instructions to the query runner.

Exploration Workflows & User Personas​

Ravioli adapts to different analytical needs and technical skill levels through three main workflows:

1. Quick Insights (Basic Users)​

Designed for users who want immediate, hassle-free answers from a single file or data source.

  • Target Persona: Operations managers, product owners, and business users who need rapid insights without writing queries.
  • Key Features: Conversational analytics, automated statistical profiling, and instant data quality checks.
  • Learn more: Quick Insights

2. Custom Notebooks (Data Analysts & Scientists)​

An advanced, cell-based interface that provides precise, granular control over data processing and documentation.

  • Target Persona: Data analysts, data scientists, and engineers who are highly comfortable writing code and queries.
  • Key Features: A flexible notebook layout supporting SQL cells, Python cells, AI cells, and Markdown cells for complete customization.
  • Learn more: Custom Notebooks

3. Deep Dives (AI-Guided Analytics) Upcoming​

An intelligent, automated workflow that bridges the gap between technical and non-technical team members.

  • Target Persona: Basic users (e.g., ops managers) who want the depth and rigor of advanced data analysis without having to write SQL or Python themselves.
  • Key Features: AI-generated multi-step analytics plans modeled after Custom Notebook structures. The agent autonomous runs queries, self-corrects on failure, and builds charts to deliver a comprehensive dashboard.
  • Learn more: Deep Dives