Widget Analytics Domain

Overview#

An analyst is three weeks into a complex fraud investigation and suspects the pattern of transactions clusters around specific weekday windows. Instead of writing a query or asking a data engineer, she types "show me transaction volume by day of week for investigation 4412" into the analytics interface. A bar chart appears in seconds, and the Wednesday spike is immediately obvious. She saves the query as a Smart Widget, pins it to her dashboard, and shares it with the team. That kind of immediate, conversational data access is what the Widget Analytics domain delivers.

The domain provides conversational analytics powered by AI, enabling natural language queries against investigation data. Users ask questions in plain language and receive AI-generated visualisations, proactive insights, and Smart Widgets for saved queries, with session-based context preservation for follow-up questions.

Key Features#

• Natural Language Queries: Ask questions about investigation data in plain language and receive structured results with appropriate visualisations, without needing technical query languages.

• Conversational Context: Maintain session context so follow-up questions build on previous queries, enabling iterative exploration of data through natural conversation.

• AI-Generated Visualisations: Automatically select and configure the best visualisation type (line charts, bar charts, pie charts, heatmaps, tables, scatter plots, area charts) based on the data and query intent.

• Smart Widgets: Save useful queries as reusable Smart Widgets that can be shared across the organisation, pinned to dashboards, and parameterised for flexible reuse.

• Proactive Insights: Receive AI-generated insights that surface patterns, anomalies, and notable findings from investigation data, with severity levels and confidence scores.

• Session Management: Create and manage analytics sessions with configurable duration and investigation context for organised, focused analysis.

Visualisation Types#

Mermaid Diagram#

Use Cases#

• Law Enforcement: Explore investigation data through natural language questions to quickly surface relevant patterns, relationships, and anomalies without waiting for a data analyst to write a query.

• Financial Crime: Build custom analytical dashboards by saving natural language queries as Smart Widgets that update with live data as the investigation progresses.

• Intelligence Analysis: Use proactive insights to discover unexpected patterns in investigation data, such as unusual communication bursts or geographic clustering, that warrant further analysis.

• Collaborative Investigations: Share Smart Widgets across the team so that useful analytical queries are reusable and consistently available to every investigator working on a case.

Integration#

The Widget Analytics domain enhances analytical capabilities across the platform:

• Investigation Management: Analytics sessions connect to investigation data.
• Analytics: Conversational queries complement traditional analytics dashboards.
• AI Services: Natural language processing and insight generation use AI capabilities.
• Dashboard: Smart Widgets integrate with platform dashboard interfaces.

Open Standards#

• GraphQL (June 2018 Specification): All analytical queries, mutations, and type definitions are exposed through a strongly-typed GraphQL schema, allowing clients to request exactly the fields they need for sessions, widgets, and insights.
• W3C Server-Sent Events (EventSource): Real-time query progress, visualisation metadata, proactive insights, and heartbeat signals are streamed to clients using the SSE protocol over text/event-stream, with named event types and sequence identifiers for reconnection.
• JSON Web Token (RFC 7519) with RS256: Every API request is authenticated by verifying a signed JWT bearer token; the domain enforces RS256 asymmetric signing validated against a JWKS endpoint, protecting all sessions and mutations.
• OAuth 2.0 (RFC 6749): The bearer-token authorisation framework underpins how clients present credentials; the domain's permission guards follow the OAuth 2.0 protected-resource pattern for access control on investigation-scoped queries.
• JSON (RFC 8259): All query results, visualisation configurations, proactive insight payloads, and SSE event data are serialised as JSON, forming the canonical interchange format between the service layer and consumers.
• ISO 8601: All session, query, and insight timestamps are serialised in ISO 8601 format via Python's .isoformat() method, ensuring consistent, unambiguous datetime interchange across time zones.
• OpenAPI Specification 3.x (OAS 3): The REST SSE streaming and heartbeat endpoints are documented automatically via FastAPI's OpenAPI integration, providing a machine-readable contract for the text/event-stream interface alongside the GraphQL layer.

Last Reviewed: 2026-02-09 Last Updated: 2026-04-14