Genpark Unified Marketing Analyst Skill

GenPark AI Agent Skill -- # Unified Marketing Analyst Skill

This repository contains the Unified Marketing Analyst Skill — a developer Python SDK client (marketing_analyst.py), an agent skill configuration (skill.json), and executable validation examples designed to integrate Google Ads, Google Analytics, and Merchant Center data feeds to audit campaigns, identify feed anomalies, and suggest optimal spend adjustments.

---

🚀 Capabilities

• Blended Performance Aggregator: Computes overall ad spend and store revenues, producing a holistic Blended ROAS metric.
• Merchant Center Feed Audits: Scans product feeds for critical disapprovals or warnings (e.g. missing GTIN values) that halt ad impressions.
• Budget Allocations Optimizer: Pinpoints campaigns with high budget burn but zero conversions, offering targeted actions to scale outperforming campaigns and pause failing ones.

---

🛠️ Setup & Installation

1. Install dependencies:

   pip install -r requirements.txt
   

2. Configuration: Set your API environment variables if executing requests against the live production server (otherwise, client executes in mock mode):

   • PowerShell:

     $env:ANALYST_API_KEY="your_api_key"
     

   • bash:

     export ANALYST_API_KEY="your_api_key"
     

---

💻 SDK Usage Reference

from marketing_analyst import MarketingAnalystClient

# Initialize Client (mock mode by default)
client = MarketingAnalystClient()

# Compute blended performance
perf = client.compute_blended_performance(
    google_ads=[{"cost": 100.00, "conversions": 3}],
    analytics=[{"revenue": 450.00}]
)
print(f"Blended ROAS: {perf['blended_roas']}x")

# Check Merchant Center feeds
audit = client.audit_merchant_center([
    {"title": "Sweater", "issue_severity": "disapproved", "issue_description": "No GTIN"}
])
print(f"Alerts: {audit['anomaly_alerts']}")

---

📜 License

This project is licensed under the MIT License.