Back to Documentation

Analytics Gateway Guide

The Analytics Gateway transforms your Gold layer medallion data into production-ready analytics workflows. Generate sophisticated data science pipelines for platforms like Metaflow, Airflow, and Databricks.

Table of Contents

Quick StartPlatform Preparation WizardMetaflow IntegrationTemplate LibraryMulti-Context AnalyticsFlow ManagementGenerated Code StructureExecution & Results

Quick Start Guide

1

Access Analytics Gateway

# Start JOEY
./dev.sh
# Navigate to Analytics Gateway
open http://localhost:6100

Launch JOEY and navigate to the Analytics Gateway tab in the UI

2

Create Your First Flow

  • Click "Create New Flow"
  • Select target platform (e.g., Metaflow)
  • Choose Gold layer context
  • Pick analysis template

Use the Platform Preparation Wizard to configure your workflow

3

Execute Generated Code

# Navigate to flows directory
cd ~/.joey-data/flows
# Execute the flow
python your_flow_name.py run

Run your generated Metaflow workflow and view comprehensive results

Platform Preparation Wizard

The wizard guides you through platform selection, context configuration, and template customization to generate the perfect analytics workflow for your needs.

Supported Platforms

Metaflow

Netflix's data science framework (Recommended)

Full Support

Airflow

Apache workflow orchestration

Full Support

Databricks

Unified analytics platform

Full Support

Kubeflow

ML workflows on Kubernetes

Full Support

Amazon Bedrock

Foundation model workflows

Full Support

Power BI

Business intelligence dashboards

Partial Support

Wizard Steps

1

Platform Selection

Choose your target analytics platform

2

Context Selection

Single or multi-context data analysis

3

Template Selection

Pick from filtered, compatible templates

4

Configuration

Customize analysis parameters

5

Preview & Generate

Review and generate your workflow

Metaflow Integration

JOEY generates production-ready Metaflow workflows that load data directly from your Gold layer, perform sophisticated analytics, and scale automatically from development to production.

Why Metaflow + JOEY?

Load data directly from Gold layer medallion architecture
Perform sophisticated analytics using enterprise algorithms
Execute in parallel for optimal performance
Generate business reports in multiple formats
Scale automatically from development to production
Integrate seamlessly with existing data infrastructure

Metaflow Features

  • • Data versioning and artifact storage
  • • Dependency management and environment isolation
  • • Scaling from laptop to cloud (AWS Batch, Kubernetes)
  • • Experiment tracking and reproducibility
  • • Built-in parallelization and step orchestration

Generated Flow Structure

# Production-ready analytics workflow generated by JOEY
from metaflow import FlowSpec, step, Parameter, batch
class
YourAnalyticsFlow
(
FlowSpec
):
# Configurable parameters
data_path =
Parameter
(
'data_path'
)
quality_threshold =
Parameter
(
'quality_threshold'
)
@step
def
start
(
self
):
# Load Gold layer data
self
.next(
self
.analyze_quality,
self
.analyze_patterns)
@batch
(cpu=
2
, memory=
4000
)
@step
def
analyze_quality
(
self
):
# Comprehensive data quality assessment
@step
def
end
(
self
):
# Generate reports and artifacts

Template Library

Pre-built analytics templates for common data science workflows. Each template is optimized for specific use cases and automatically adapts to your data structure.

Data Quality Check

Comprehensive data validation and quality reporting

Single Context

Best For:

  • Data governance
  • Quality reporting
  • Data validation

Statistical Analysis

Descriptive statistics, correlations, and distributions

Multi-Context

Best For:

  • Data exploration
  • Statistical profiling
  • Correlation analysis

Feature Engineering

ML-ready feature preparation and transformation

Multi-Context

Best For:

  • Machine learning
  • Feature preparation
  • Data transformation

Comparison Analysis

Cross-context comparison and correlation analysis

Multi-Context

Best For:

  • Business intelligence
  • Benchmarking
  • Comparative analysis

Time Series Forecast

Predictive analytics for time-based data

Single Context

Best For:

  • Forecasting
  • Trend analysis
  • Predictive modeling

Aggregation Report

Business summary analytics and KPI calculation

Multi-Context

Best For:

  • Executive dashboards
  • KPI tracking
  • Business reporting

Execution & Results

Execute your generated workflows and access comprehensive results including executive summaries, detailed JSON reports, and interactive Metaflow artifacts.

Execution Commands

Basic Execution

# Run with default parameters
python your_flow.py run
# Run with custom parameters
python your_flow.py run --quality_threshold 0.90

Viewing Results

# Show all runs
python your_flow.py show
# View generated HTML cards
python your_flow.py card view

Cloud Execution

# Run on AWS Batch
python your_flow.py --with batch run
# Custom compute resources
python your_flow.py --with batch:cpu=4,memory=8000 run

Expected Results

Console Output

🔍 Starting Enterprise Data Quality Assessment
📊 Data source: ~/.joey-data/gold/context/current_data.json
🎯 Quality threshold: 95.0%
📈 Dataset structure: 1,000 rows × 15 columns
✅ Completeness: 94.2%
✅ Uniqueness: 98.7%
✅ Validity: 96.1%
✅ Consistency: 99.2%
🎯 Overall Quality Score: 97.1% (EXCELLENT)
Executive summary reports
Detailed JSON analytics results
Interactive Metaflow artifacts
Downloadable visualizations

Ready to Build Your First Workflow?

Follow the quick start guide and create production-ready analytics workflows in minutes.

Start MetaFlow GuideView All Docs