Skip to content

README.md

Latest commit

 

History

History
167 lines (138 loc) · 7.29 KB

README.md

File metadata and controls

167 lines (138 loc) · 7.29 KB

Data Analytics

AI-powered multi-agent system for capital markets data analysis -- data exploration, statistical analysis, and actionable insight generation.

Overview

The Data Analytics use case coordinates three specialist agents through a DataAnalyticsOrchestrator to help capital markets analysts derive insights faster from financial datasets. It profiles data quality, performs rigorous statistical analysis, and translates findings into business-relevant conclusions with visualization recommendations.

Business Value

  • Accelerated analysis -- Parallel data exploration, statistics, and insight generation compress days of analyst work
  • Data quality assurance -- Automated profiling catches completeness, consistency, and accuracy issues before decisions are made
  • Statistical rigor -- Hypothesis testing and regression modeling with significance levels, effect sizes, and confidence intervals
  • Actionable output -- Insights tagged with confidence levels and paired with specific visualization recommendations
  • Repeatable methodology -- Consistent analytical framework across datasets and teams

Architecture

graph TB
    Request["Client Request"] --> Runtime["AgentCore Runtime"]
    Runtime --> Orchestrator["Data Analytics<br/>Orchestrator"]
    Orchestrator --> Agent1["Data Explorer<br/><small>Explores datasets, profiles quality,<br/>identifies patterns and anomalies</small>"]
    Orchestrator --> Agent2["Statistical Analyst<br/><small>Performs hypothesis testing,<br/>regression modeling</small>"]
    Orchestrator --> Agent3["Insight Generator<br/><small>Generates actionable insights,<br/>visualization recommendations</small>"]
    Agent1 --> Bedrock["Amazon Bedrock<br/>(Claude)"]
    Agent2 --> Bedrock
    Agent3 --> Bedrock
    Agent1 --> S3["S3 Sample Data<br/>(Analytics Datasets)"]
    Agent2 --> S3
    Agent1 --> Synthesis["Result Synthesis"]
    Agent2 --> Synthesis
    Agent3 --> Synthesis
    Synthesis --> Response["Response"]
Loading

Directory Structure

use_cases/data_analytics/
├── README.md
└── src/
    ├── __init__.py                              # Framework router + registry
    ├── strands/
    │   ├── __init__.py
    │   ├── config.py                            # DataAnalyticsSettings
    │   ├── models.py                            # AnalyticsRequest / AnalyticsResponse
    │   ├── orchestrator.py                      # DataAnalyticsOrchestrator
    │   └── agents/
    │       ├── __init__.py
    │       ├── data_explorer.py
    │       ├── statistical_analyst.py
    │       └── insight_generator.py
    └── langchain_langgraph/
        ├── __init__.py
        ├── config.py
        ├── models.py
        ├── orchestrator.py
        └── agents/
            ├── __init__.py
            ├── data_explorer.py
            ├── statistical_analyst.py
            └── insight_generator.py

Agentic Design

The DataAnalyticsOrchestrator extends StrandsOrchestrator and uses a parallel fan-out / synthesize pattern:

  1. Fan-out -- For full assessments, all three agents run via asyncio.gather, each independently retrieving entity data from S3.
  2. Targeted modes -- data_exploration runs the explorer alone; statistical_analysis pairs explorer + statistician; insight_generation pairs statistician + insight generator.
  3. Synthesis -- Agent results are assembled into section-labeled markdown and fed to the orchestrator LLM, which produces a structured JSON summary with data quality classification, patterns, statistical findings, and recommendations.

Agents

Data Explorer

  • Role: Explores datasets to understand structure, distributions, and quality; detects patterns, correlations, and anomalies
  • Data: Entity profile from S3 (data_type='profile')
  • Produces: Data quality assessment (completeness, consistency, accuracy), patterns and trends, outlier detection, recommended analytical approaches
  • Tool: s3_retriever_tool

Statistical Analyst

  • Role: Performs rigorous statistical analysis including hypothesis testing, regression modeling, and significance evaluation
  • Data: Entity profile from S3
  • Produces: Test results with p-values, regression model summaries, effect sizes and confidence intervals, model diagnostics
  • Tool: s3_retriever_tool

Insight Generator

  • Role: Translates analytical findings into business-relevant conclusions with confidence-scored insights
  • Data: Entity profile from S3
  • Produces: Key insights with confidence levels, business implications, visualization suggestions (charts, dashboards, heatmaps), areas for further investigation
  • Tool: s3_retriever_tool

Data & Tools

Resource Description
s3_retriever_tool Retrieves dataset profiles and data from S3
S3 path data/samples/data_analytics/{entity_id}/profile.json

Request / Response

AnalyticsRequest

Field Type Description
entity_id str Dataset/entity identifier (e.g., ASSET001)
assessment_type AssessmentType full, data_exploration, statistical_analysis, insight_generation
additional_context str | None Optional context

AnalyticsResponse

Field Type Description
entity_id str Dataset/entity identifier
analytics_id str Unique assessment UUID
timestamp datetime Assessment timestamp
analytics_detail AnalyticsDetail | None Data quality, patterns, statistical findings, visualization suggestions, coverage %
recommendations list[str] Analytical recommendations
summary str Executive summary
raw_analysis dict Raw output from each agent

Example Request:

{
  "entity_id": "ASSET001",
  "assessment_type": "full",
  "additional_context": "Focus on volatility patterns"
}

Example Response:

{
  "entity_id": "ASSET001",
  "analytics_id": "uuid",
  "timestamp": "2026-03-25T00:00:00Z",
  "analytics_detail": {
    "data_quality": "high",
    "insight_confidence": "high",
    "patterns_identified": ["Mean-reverting volatility pattern in tech sector"],
    "statistical_findings": ["Significant correlation between VIX and sector returns (p<0.01)"],
    "visualization_suggestions": ["Rolling volatility heatmap", "Correlation matrix"],
    "data_coverage_pct": 98.5
  },
  "recommendations": ["Monitor volatility clustering for position sizing"],
  "summary": "High-quality dataset with strong statistical patterns identified..."
}

Quick Start

USE_CASE_ID=data_analytics FRAMEWORK=strands AWS_REGION=us-east-1 \
  ./applications/fsi_foundry/scripts/deploy/full/deploy_agentcore.sh

Sample Data

Entity ID Description
ASSET001 US Equity Sector Performance Dataset -- Technology sector, daily data 2020-2024

Related Documentation