An autonomous AI agent that monitors competitor car rental pricing across 10 major US airports, integrates real-time demand signals, and generates AI-powered pricing recommendations — replacing hours of manual analyst work with a 15-minute automated pipeline.
ABG's Revenue Management team manually checks competitor prices across airports every day — opening Expedia, Kayak, Hertz.com, building Excel reports, making gut calls. It takes hours across a 20-person team.
This agent does it automatically in 15 minutes.
┌─────────────────────────────────────────────────────┐
│ PIPELINE (runs every 6 hours) │
└──────────┬──────────────┬──────────────┬────────────┘
│ │ │
┌───────▼──────┐ ┌────▼──────┐ ┌────▼──────────┐
│ Agent 1 │ │ Agent 2 │ │ Agent 3 │
│ Competitor │ │ Demand │ │ AI Pricing │
│ Prices │ │ Signals │ │ Recs (Groq) │
│ │ │ │ │ │
│ 8 competitors│ │ • Flights │ │ • llama-3.3 │
│ 10 airports │ │ • Events │ │ • 70b model │
│ All categories│ │ • Weather │ │ • Urgency │
└──────┬───────┘ └────┬──────┘ └────┬──────────┘
└──────────────┴──────────────┘
│
┌──────▼──────┐
│ SQLite │
│ pricing.db │
└──────┬──────┘
│
┌──────▼──────┐
│ Streamlit │
│ Dashboard │
└─────────────┘
5-tab Streamlit dashboard showing:
| Tab | What it shows |
|---|---|
| 📊 Price Heatmap | Competitor rates by category across all airports |
| 🤖 AI Recommendations | Groq-generated pricing actions sorted by urgency |
| Competitor price changes ≥20% flagged automatically | |
| 📈 Competitor Analysis | Bar charts, box plots, rate breakdowns |
| 🗺️ Network Overview | US map of economy rates across all 10 airports |
| Component | Technology | Cost |
|---|---|---|
| AI Recommendations | Groq (llama-3.3-70b-versatile) | Free |
| Flight Demand | OpenSky Network API | Free |
| Events Data | Ticketmaster Discovery API | Free |
| Weather Forecasts | Open-Meteo API | Free |
| Dashboard | Streamlit | Free |
| Database | SQLite | Free |
| Scheduler | GitHub Actions | Free |
| Language | Python 3.11 | Free |
Total monthly cost: $0
git clone https://github.com/Mkp-7/Revenue-Management-Agent.git
cd Revenue-Management-Agent
pip install -r requirements.txtcp .env.example .envEdit .env:
GROQ_API_KEY=your_key # console.groq.com — free
TICKETMASTER_API_KEY=your_key # developer.ticketmaster.com — free
OPENSKY_USERNAME= # optional — opensky-network.org
OPENSKY_PASSWORD= # optionalpython orchestrator.py --oncestreamlit run dashboard/app.pyRevenue-Management-Agent/
├── agents/
│ ├── agent1_scraper.py # Competitor price engine
│ ├── agent2_demand.py # Flight + event + weather signals
│ └── agent3_recommendations.py # Groq AI pricing recommendations
├── config/
│ ├── airports.py # 50 US airports with ICAO codes
│ ├── database.py # SQLite schema
│ └── settings.py # Shared config (airports list etc.)
├── dashboard/
│ └── app.py # Streamlit 5-tab dashboard
├── .github/workflows/
│ └── pipeline.yml # GitHub Actions — runs every 6h
├── orchestrator.py # Master pipeline runner
├── requirements.txt
└── .env.example
| Code | City | Code | City |
|---|---|---|---|
| ATL | Atlanta | JFK | New York |
| LAX | Los Angeles | SFO | San Francisco |
| ORD | Chicago | LAS | Las Vegas |
| DFW | Dallas | MCO | Orlando |
| DEN | Denver | MIA | Miami |
Hertz · Avis · Enterprise · National · Alamo · Budget · Dollar · Thrifty
Groq (llama-3.3-70b-versatile) analyzes competitor prices + demand signals and outputs:
[HIGH] ECONOMY ↑ +$8/day
Raise economy rates by $8 at ATL — Braves game (45,000 attendance) driving demand spike.
[CRITICAL] SUV ↑ +$22/day
Raise SUV rates 20% at MIA — Hurricane warning forecast drives last-minute rental surge.
[MEDIUM] COMPACT ↓ −$5/day
Cut compact rates at ORD — Budget undercutting market by 18%, volume risk.
Pipeline auto-runs every 6 hours on GitHub's servers — free forever.
To trigger manually: Actions → ABG Pricing Intelligence Pipeline → Run workflow
Built to demonstrate how autonomous AI agents can replace manual revenue management workflows in the car rental industry. It demonstrates:
- Autonomous multi-agent AI architecture
- Real-time data pipeline design
- Revenue management domain knowledge
- Full-stack development (Python, SQLite, Streamlit, GitHub Actions)
- Production-ready code quality
"The same workflow analysts do manually in 3 hours, this does in 15 minutes."