# Getting started
<a name="data-mesh-getting-started"></a>

## Platform Foundation Deployment (Optional)
<a name="platform-foundation-deployment-optional"></a>

Deploy SageMaker Unified Studio for centralized data governance.

## Prerequisites
<a name="prerequisites"></a>
+ Customer 360 and/or Predictive Maintenance already deployed
+ AWS IAM Identity Center configured
+ VPC with private subnets

## Deploy SageMaker Unified Studio
<a name="deploy-sagemaker-unified-studio"></a>

```
# Navigate to platform foundation directory
cd automotive-data-platform-on-aws/platform-foundation

# Install dependencies
npm install

# Deploy
cdk deploy PlatformFoundationStack
```

 **Deployment time**: 60 minutes

 **What gets deployed**:
+ SageMaker Unified Studio domain
+ Amazon DataZone domain
+ Lake Formation resource shares
+ IAM roles for cross-domain access

## Register Data Products
<a name="register-data-products"></a>

 **Register Customer 360**:

1. Open DataZone console

1. Create project: "Customer Analytics"

1. Publish data product:
   + Name: `customer-360-analytics` 
   + Source: Glue Data Catalog `cx_analytics` 
   + Tables: All 11 tables
   + Access: Request-based approval

 **Register Predictive Maintenance**:

1. Create project: "Vehicle Intelligence"

1. Publish data product:
   + Name: `predictive-maintenance` 
   + Source: Glue Data Catalog `mmt_predictive_maintenance` 
   + Tables: All prediction tables
   + Access: Automated approval for anonymized data

## Configure Cross-Domain Access
<a name="configure-cross-domain-access"></a>

```
# Grant Customer Analytics project access to Predictive Maintenance data
aws datazone create-subscription-grant \
  --domain-identifier dzd-... \
  --granted-entity '{"listing":{"identifier":"lst-...","revision":"1"}}' \
  --subscription-target-identifier st-... \
  --region us-east-1
```

## Test Cross-Domain Query
<a name="test-cross-domain-query"></a>

```
-- Query combining Customer 360 and Predictive Maintenance data
SELECT
  c.customer_id,
  c.health_score,
  c.nps,
  p.risk_level,
  p.days_to_failure,
```