A full-stack data analytics solution for monitoring, forecasting, and automating digital ad performance metrics — built with Python, Streamlit, and Prophet.

The Ad Intelligence Suite empowers marketers and analysts to:

• Track key ad metrics like eCPM, CTR, Clicks, Revenue, and Impressions
• Detect anomalies in campaign performance (e.g. drops/spikes)
• Forecast future ad performance using time series modeling
• Automatically send alerts via email and Slack
• Explore data via an interactive dashboard

This project combines automation, forecasting, anomaly detection, and visualization into a real-world, production-ready data product.

ad-intelligence-suite/
├── alerts/                 # Email + Slack alert scripts
├── app/                    # Utilities (helpers, configs)
├── dashboard/              # Streamlit dashboard
├── forecast/               # Forecasting scripts
├── images/                 # 📸 All screenshots/images
├── data/                   # Raw datasets (CSV)
├── reports/                # Outputs (forecasts, anomalies, HTML)
├── notebooks/              # EDA, anomaly detection, modeling
├── .env                    # Email/Slack credentials (excluded in .gitignore)
└── README.md

• Data profiling
• Missing value checks
• Distribution & trend plots across eCPM, CTR, Revenue, and more

• Identify abnormal CTR, eCPM, Clicks, and Impressions
• Output a clean anomaly_summary.csv report
• Alert stakeholders automatically

• Prophet-based time series modeling
• Auto-updated forecast file (forecast_ecpm.csv)
• Weekly summary messages (e.g. “eCPM down 8.4%”)

• Anomaly alerts sent instantly
• Weekly forecast alerts via Yahoo Mail + Slack
• Uses .env for secure credentials

• Metric switcher (CTR, Clicks, Revenue, etc.)
• Highlight anomalies directly on charts
• Compare date ranges (WoW, MoM)
• Forecast lines and historical trends displayed together

• Fully autonomous with Windows Task Scheduler
• No need to run manually

Email & Slack Alerts	Forecast Visuals

1. Clone the repo

   git clone https://github.com/HenryMorganDibie/ad-intelligence-suite.git
   cd ad-intelligence-suite
   
   

2. Create a virtual environment

   python -m venv .venv
   .venv\Scripts\activate
   pip install -r requirements.txt
   
   

3. Set up your .env

   EMAIL_SENDER=your_yahoo_email
   EMAIL_PASSWORD=your_yahoo_app_password
   EMAIL_RECEIVER=recipient@example.com
   SLACK_WEBHOOK=https://hooks.slack.com/...
   
   

4. Run components

   - Run EDA: eda/eda.ipynb
   
   - Forecast: python forecast/generate_forecast_and_anomalies.py
   
   - Anomaly Alerts: python alerts/send_alerts.py
   
   - Forecast Alerts: python forecast/send_forecast_alerts.py
   
   - Dashboard: streamlit run dashboard/dashboard.py
   
   

💡 Inspiration This project simulates the kind of real-time, data-driven alerting system used by performance marketing teams to optimize digital ad spend and delivery in real-time. It was built to demonstrate full-stack data capabilities in a real-world setting.

📫 Contact Henry C. Dibie henrymorgan273@yahoo.com

Give this repo a ⭐ if you find it useful or inspiring!