Analytics for the Engage Project

This includes:

• dbt models for transforming healthcare data
• Metrics and indicators for program monitoring
• Claude Code skill for managing indicators

Quick Start

Prerequisites

• PostgreSQL database with Airbyte-synced FHIR data
• Python 3.10+ with uv package manager
• dbt-core and dbt-postgres

Running dbt

cd dbt
source .envrc  # or set DBT_HOST, DBT_PORT, DBT_USER, DBT_PASSWORD, DBT_DATABASE

# Run all models
uv run dbt run --profiles-dir .

# Run metrics only
uv run dbt run --profiles-dir . --select +fct_metrics_long

Claude Code Skill

This repo includes a Claude Code skill for managing Engage indicators and metrics.

Installation

Add this marketplace to Claude Code:

onaio/engage-analytics

Install the engage-analytics skill

claude mcp add-marketplace onaio/engage-analytics

Then enable the plugin:

claude mcp enable engage-indicators@engage-analytics

Or install directly from GitHub:

claude mcp add-marketplace https://github.com/onaio/engage-analytics

What the Skill Helps With

1. Understand existing indicators - Query metrics, see indicator mappings
2. Create new indicators - Guided workflow asking the right questions
3. Add new questionnaires - Update metadata, generate models
4. Update anonymization - Manage PII field settings
5. Test and validate - Verify metric logic

Usage

Once installed, the skill triggers automatically when you ask about:

• Indicator creation or metrics
• Questionnaire data models
• PHQ-9/GAD-7 scores
• mwTool eligibility
• dbt model development for Engage

Example prompts:

• "Show me the current indicators"
• "Help me create a new indicator for client retention"
• "Add the new intake form questionnaire"
• "What metrics are available for IPC?"

Documentation

• Metrics Documentation - Full indicator and metric reference
• Data Analyst Guide - System architecture and workflow guide
• dbt Project Guide - dbt project structure and commands