Author: Rijo Mathew John
Degree: MSc Data Analytics | Decision Intelligence & Operations Analytics
Live Demo: 🎯 Launch Interactive App
Built With: Python • Streamlit • Plotly • scikit-learn • ENTSO-E • Meteostat
Ireland’s electricity grid is in constant tension — balancing Affordability, Reliability, and Sustainability.
This project builds a Decision Intelligence framework that connects:
- Hourly generation & demand (ENTSO-E)
- Weather impacts (Meteostat)
- Market price behavior
- Scenario simulations for energy policy & operations
It’s not just visualization — it’s a working model that tests decisions before implementing them.
Ireland’s grid shows high price sensitivity to renewable volatility.
When wind generation drops, stress hours spike and prices surge disproportionately.
This project explores:
- How weather and renewables shape market volatility
- How system stress relates to cost and reliability
- What strategies can stabilize the grid without raising costs
To design a data-driven decision support tool that quantifies the trade-offs between:
- 💶 Price stability
- ⚙️ Grid reliability
- 🌱 Renewable consistency
And to simulate how operational policies — like increasing wind, smoothing renewables, or managing demand — affect these KPIs in real time.
- Combine hourly generation, demand, weather, and price data for one week (Oct 18–25 2025).
- Find patterns — where do prices jump and stress hours occur?
- Train a model that links price to real system variables (renewable share, load, volatility).
- Build sliders to simulate “what-if” scenarios — e.g., What if wind output grows 20%?
- Observe how that shifts the grid’s performance metrics.
- Data Sources:
- ENTSO-E Transparency Platform — Load, Generation, Market Prices
- Meteostat — Weather (wind, temperature, sunshine)
- Processing:
- Python (Pandas, NumPy) with hourly normalization
- Star-schema “mart” of load, generation, renewables, weather, price
- Modeling:
- Ridge Regression →
price ~ ren_share + stress + load_scaled (+ weather) - Auto-handling of NaN/Inf + feature scaling (scikit-learn Pipeline)
- Ridge Regression →
- Simulation Engine:
- Dynamic parameter adjustment via Streamlit sliders
- KPI comparison (avg price, stress %, RSD of renewables)
- Real-time Plotly dashboards
| Insight | Observation |
|---|---|
| 💨 Wind volatility | High variability (RSD > 0.25) directly drives price spikes |
| ⚙️ Stress hours | When generation < demand, stress > 10 % → price surges |
| 🌅 Peak hours | 17:00 – 21:00 hrs remain high-stress even on stable days |
| 💶 Price-renewable link | Prices fall almost linearly with renewable % > 60 % |
Gap Identified:
The grid is too reactive — a small renewable dip triggers large market swings.
→ Ireland needs more resilience, stability, and foresight.
| Action | Impact |
|---|---|
| Increase Wind Penetration (+20%) | Reduces avg price / stress hours |
| Stabilize Renewables (–25% RSD) | Smoother supply curve, better reliability |
| Prepare for Demand Growth (+10%) | Requires flexible storage / responsive generation |
| Hybrid Strategy | Combines cost reduction + resilience gain |
The dashboards below form the analytical layer that feeds into the Decision Intelligence Simulator.
They summarize how Ireland’s energy system performs across affordability, reliability, and renewable stability.
Load vs Generation, Price evolution, and Renewable share dynamics.
Visualizing stress vs non-stress pricing behavior and demand sensitivity.
How wind efficiency and sunshine patterns influence renewable variability.
Live Ridge Regression–based scenario simulator connecting data → prediction → insight.
🎯 Try it yourself: Launch Interactive App →
Interactive Controls
- Adjust
Wind,Demand, andStabilitysliders - Observe real-time KPI shifts and price trajectories
KPIs
- 💶 Average Price
- ⚙️ Stress Hours
- 🌱 Renewable Stability (RSD)
Technically
Model: Ridge Regression (scikit-learn)
Features: ren_share, stress, load_scaled, wind_speed, sunshine_fraction
Data Window: Oct 18–25 2025 (hourly)
---
## 📈 8. Key Results
| Metric | Baseline | Simulated (Hybrid) | Δ Change |
| :------------------------ | :-------------- | :----------------- | :----------------- |
| Average Price (€/MWh) | ↓ From live app | ↓ | Cost reduction |
| Stress Hours (%) | ↓ | ↓ | Higher reliability |
| Renewable Stability (RSD) | ↓ | ↓ | Smoother operation |
📊 The **hybrid scenario (Wind + Stability)** gave the **best multi-objective balance**.
---
## 🧱 9. Architecture Overview
```text
ENTSO-E Data ─┬─> Load ─┐
├─> Generation ─┐
Meteostat Data ─┘ │
├─> Data Mart (hourly)
│
├─> EDA + KPI Engine
│
└─> Simulation (Ridge Model)
↓
Streamlit Decision Layer
| Layer | Tools Used |
|---|---|
| Data Processing | Python (pandas, numpy) |
| Modeling | scikit-learn (Ridge Regression) |
| Visualization | Plotly Express + Graph Objects |
| App Layer | Streamlit (Dark Theme UI) |
| Data Sources | ENTSO-E • Meteostat |
| Deployment | Streamlit Cloud + GitHub |
Decision Intelligence for Energy Balance — not just a dashboard.
This project fuses operational data analytics with strategic simulation, quantifying trade-offs between:
- Cost 🪙
- Reliability ⚙️
- Sustainability 🌱
It converts static energy reporting into a decision-making tool — a glimpse of how future smart grids will be managed.
# clone the repository
git clone https://github.com/rijomj008-create/Ireland-Energy-Simulator.git
cd Ireland-Energy-Simulator
# install dependencies
pip install -r requirements.txt
# run locally
streamlit run app.pyThen open http://localhost:8501
Rijo Mathew John 📍 Dublin, Ireland 🎓 MSc Data Analytics — Dublin Business School 💼 Decision Intelligence | Operations Analytics | Energy Systems 📧 rijomj008@gmail.com 🔗 LinkedIn →
MIT License — You’re welcome to reuse with attribution.
⭐ If you found this project useful, please star ⭐ the repository — it helps others discover Decision Intelligence for Energy.
---
### ✅ How this aligns with your vision
- **Narrative-first** flow (Problem → Goal → Methodology → Simulator → Findings → CTA).
- Mix of **layman & analytical tone** for recruiters and domain experts alike.
- **Polished formatting** with icons, tables, and clickable demo link.
- **Future-proof** — you can reuse this layout for other Decision Intelligence projects.
Would you like me to include a matching `config.toml` theme section (so the app dark mode matches this README’s aesthetic)?