PJM Peak Analysis Tool

Analyzes peak demand hours and peak pricing hours across all PJM zones, identifies overlap patterns, and provides a 7-day forecast.

Features

• Historical Analysis: Identifies top 5 peak load hours and top 5 peak price hours per month for each PJM zone
• Overlap Analysis: Quantifies how often peak load and peak price hours coincide
• 7-Day Forecast: Predicts likely peak hours based on historical patterns and PJM load forecasts
• Reporting: Generates daily markdown reports
• Dashboard: Interactive visualization with Dash/Plotly

Quick Start

1. Setup Environment

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

2. Configure API Credentials

# Copy example env file
cp .env.example .env

# Edit .env and add your PJM subscription key
# Get your key from: https://dataminer2.pjm.com/

3. Run the Analysis

# Full daily pipeline
python main.py

# Options:
python main.py --fetch-only     # Only fetch new data
python main.py --report-only    # Generate report from cached data
python main.py --zone PECO      # Run for specific zone only
python main.py --mock           # Use mock data (no API key required)

4. Launch Dashboard

# Streamlit (recommended)
streamlit run dashboard/streamlit_app.py
# Open http://localhost:8501 in your browser

# Or Dash
python dashboard/app.py
# Open http://localhost:8050 in your browser

Project Structure

pjm-peak-analysis/
├── config/
│   └── settings.py          # Configuration and zone mappings
├── data/
│   ├── api_client.py        # PJM Data Miner API wrapper
│   └── mock_data.py         # Mock data generator for testing
├── analysis/
│   └── peaks.py             # Peak identification and overlap analysis
├── forecast/
│   └── price_predictor.py   # Historical pattern-based price forecasting
├── dashboard/
│   ├── streamlit_app.py     # Streamlit interactive dashboard
│   └── app.py               # Dash interactive dashboard
├── outputs/
│   ├── reports/             # Generated markdown reports
│   ├── charts/              # Generated visualizations
│   └── data/                # CSV/Parquet data exports
├── main.py                  # Main execution script
├── requirements.txt
└── PROJECT_SPEC.md          # Detailed specification document

Data Feeds Used

Feed	Purpose
hrl_load_metered	Hourly metered load by zone
rt_hrl_lmps	Real-time hourly LMPs
da_hrl_lmps	Day-ahead hourly LMPs
load_frcstd_7_day	7-day load forecast

Analysis Methodology

Peak Identification

• Ranks all hours in each month by load (MW) and price ($/MWh)
• Selects top 5 hours for each metric per zone

Overlap Metrics

• Exact overlap: Hours appearing in both top-5 lists
• Near overlap: Peak hours within ±1 hour of each other
• Hourly correlation: Pearson correlation between load and price

Forecast Model

• Matches current conditions to similar historical hours
• Weights by recency and load similarity
• Provides confidence intervals based on historical variance

Zones Covered

All 19 PJM transmission zones plus RTO aggregate:

Zone	Utility
AEP	American Electric Power
APS	Allegheny Power
ATSI	FirstEnergy Ohio
BGE	Baltimore Gas & Electric
COMED	Commonwealth Edison
DAY	Dayton Power & Light
DEOK	Duke Energy Ohio/Kentucky
DOM	Dominion Virginia
DPL	Delmarva Power
DUQ	Duquesne Light
EKPC	East Kentucky Power Coop
JCPL	Jersey Central P&L
METED	Met-Ed
PECO	PECO Energy
PENELEC	Penelec
PEPCO	Potomac Electric Power
PPL	PPL Electric
PSEG	Public Service E&G
RECO	Rockland Electric

API Rate Limits

• PJM Members: 600 requests/minute
• Non-members: 6 requests/minute

The API client includes automatic rate limiting.

License

Internal use only. Data usage subject to PJM Data Miner terms of service.

Streamlit Cloud Deployment