Feeder Manager, PHED · Data Scientist · Power Distribution Systems
BE Electrical · MBA · MSc Financial Engineering (WorldQuant)
profile = {
"name" : "Opiribo Olaniyo",
"role" : "Feeder Manager — Port Harcourt Electricity Distribution Plc (PHED)",
"education" : ["BE Electrical", "MBA (UoPeople)", "MSc Financial Engineering (WorldQuant)"],
"location" : "Port Harcourt, Rivers State, Nigeria",
# Real outcomes from PHED operations
"delivered" : {
"operational_efficiency" : "+15% via structured KPI monitoring",
"reporting_accuracy" : "+30% through process redesign & data validation",
"stockout_reduction" : "-22% via demand forecasting & cross-team planning",
},
"portfolio" : "ML pipelines built around the operational challenges I manage daily at PHED",
"open_to" : [
"EPC Oil & Gas — Technical / Interface Engineering",
"Project Planning & Control (Primavera P6 / MS Project)",
"O&M Coordination",
"Revenue Assurance & Commercial Loss Analytics",
"Data Science roles in Energy / Power Sector",
],
}These projects are not academic exercises — they are built around feeder loss, theft patterns, demand behaviour, and transformer reliability challenges I work with daily as Feeder Manager at PHED.
| # | Project | Model | Result |
|---|---|---|---|
| 🔌 | Feeder Loss Prediction | XGBoost Classifier + Regressor | 90% accuracy · 87% CV · MAE 8.5% |
| 🔦 | Energy Theft Detection | XGBoost + Random Forest | AUC 1.0 · Precision-Recall analysis |
| 📈 | Electricity Demand Forecast | XGBoost Time-Series | MAPE 4.1% · Directional Acc. 82% |
| ⚙️ | Transformer Failure Prediction | XGBoost + Random Forest | AUC 0.89 · 83% accuracy |
Built directly from my experience managing technical and commercial losses across PHED distribution zones. Flags feeders exceeding 20% loss threshold and estimates exact loss percentage for maintenance dispatch prioritisation.
Classification:
Normal Loss (≤20%) → Precision: 0.93 Recall: 0.94 F1: 0.94
High Loss (>20%) → Precision: 0.80 Recall: 0.76 F1: 0.78
Overall Accuracy → 90.3% | 5-Fold CV: 87.0%
Regression:
MAE → 8.5% RMSE → 10.1%
Business Impact:
One corrected feeder → ₦35M+/year
Network-wide 20% cut → ₦1.1B+ annually
| Area | Result | Method |
|---|---|---|
| Operational Efficiency | +15% | KPI monitoring & planning frameworks |
| Reporting Accuracy | +30% | Process redesign & data validation |
| Stock-out Reduction | -22% | Demand forecasting & cross-team planning |
🔵 EPC Oil & Gas — Technical / Interface Engineering
🔵 Project Planning — Primavera P6 / MS Project
🔵 O&M Coordination — Operations & Maintenance
🔵 Revenue Assurance — Commercial Loss Analytics
🔵 Data Science — Energy & Power Sector
📍 Port Harcourt · Warri · Lagos · Open to relocation
git clone https://github.com/opriolani/Energy_Data_Portfolio_Project
cd Energy_Data_Portfolio_Project
pip install -r requirements.txt
python data_generation/synthetic_feeder_generator.py
streamlit run app.py"Data without insight is noise. Insight without action is waste."