Skip to main content

Query and Improve your data using reasoning models and prompts.

Architecture​

QALITA Studio is an AI-powered conversational interface for data quality analysis, available directly in the Platform web application.

Key Components​

ComponentDescription
Backend AgentLangGraph-powered AI agent integrated into Platform backend
LLM ConfigurationMulti-provider AI configuration managed in Platform settings
Data ToolsDirect data access via Worker infrastructure
Frontend InterfaceIntegrated chat UI in Platform web application

Studio includes:

  • 🤖 LangGraph Agent - Intelligent workflow for data analysis and codification
  • 🔧 Data Tools - Execute queries, read data, and describe sources directly
  • 🌐 Multi-Provider LLM - Support for OpenAI, Anthropic, Mistral, Ollama, and more
  • 💬 Streaming Chat - Real-time response generation with SSE

Studio is a project supported by BPI France and is currently under development. Documentation is subject to quick improvement upon the beta release of the tool.

To learn more and stay updated, check out our blog.

🔍 Key Features​

QALITA Studio is an advanced conversational interface that allows you to interact with your data quality through language models (LLM).

Multi-Provider Artificial Intelligence​

  • Local Provider: Use Ollama to run LLM models locally (free, private)
  • Cloud Providers: Native integration with OpenAI (ChatGPT), Mistral AI, Claude (Anthropic), and Gemini (Google)
  • Azure OpenAI: Enterprise-grade Azure integration
  • Generic Provider: Connect to any OpenAI-compatible API endpoint

Intelligent Data Tools​

Studio's agent can interact directly with your data sources through the Worker infrastructure:

  • Execute SQL Queries: Run SQL queries on database sources
  • Read Data: Preview data from files (CSV, Excel, Parquet) or database tables
  • Describe Sources: Get metadata, schemas, and row counts
  • Filter & Aggregate: Apply filters and perform aggregations
  • Sample Data: Get random samples for analysis

Contextual Enrichment​

  • Issue Context: Prompts are automatically enriched with issue details (status, scope, recommendations)
  • Source Metadata: Automatic access to source configuration and schema
  • Quality Metrics: Integration of quality scores and metrics in conversations
  • Data Samples: Include data previews in the conversation context

Streaming Mode​

  • Real-Time Responses: Display model responses as they are generated
  • SSE Protocol: Server-Sent Events for smooth streaming experience
  • Graceful Fallback: Automatic fallback to chunked delivery if streaming unavailable

QALITA Platform Integration​

  • Projects & Sources: Direct access to your projects and data sources
  • Issues: Automatic context from data quality issues
  • Conversation History: Conversations stored per issue and source context
  • Multi-tenant: Full partner isolation and role-based access

studio-mockup-2.png
studio-mockup-1.png

Use Cases​

Quality Issue Investigation​

Use Studio to analyze and understand data quality issues detected by QALITA:

  • Ask for a detailed explanation of an anomaly
  • Get correction suggestions based on actual data
  • Explore patterns in your data with direct queries

Source Exploration​

Query your data sources naturally:

  • "Which columns have the most null values?"
  • "Show me a sample of records where email is missing"
  • "Explain the structure of this source"

Medical Coding Analysis​

Studio includes specialized capabilities for medical data interoperability:

  • Identify columns suitable for codification (ICD-10, SNOMED CT, LOINC)
  • Generate mapping scripts for terminology standards
  • Analyze data patterns for coding opportunities

Documentation Generation​

Automatically generate documentation about your data:

  • Table and column descriptions
  • Quality rule documentation
  • Analysis reports

Security and Privacy​

  • Role-Based Access: Studio respects Platform permissions (source:get, issue:get)
  • Partner Isolation: Complete data isolation between tenants
  • Local Mode: Ability to use Ollama for fully local processing
  • API Key Security: Credentials stored securely in Platform database

Next Steps​

  1. Quick Start - Access and configure Studio in minutes
  2. Configuration - Configure your AI providers
  3. Features - Discover all features in detail
  4. Conversation Management - Learn to manage your conversations