Operational Efficiency Scoring System

An AI-powered system that analyzes order fulfillment operations, detects bottlenecks, provides improvement recommendations, and predicts future performance.

What It Does

1. Efficiency Scoring — Scores each process step (0-100) based on cycle time, wait time, error rate, and rework rate
2. Bottleneck Detection — Identifies which steps are slowing down the entire pipeline
3. Visualizations — Creates 6 charts showing performance patterns
4. AI Recommendations — Uses Google Gemini (free) to suggest process improvements
5. Performance Prediction — Forecasts future performance using Machine Learning (Random Forest)
6. PDF Report — Generates a professional report with all findings

The Pipeline Analyzed

Order Received → Order Validation → Inventory Check → Picking & Packing
→ Quality Inspection → Shipping Prep → Dispatch & Delivery

Setup

git clone https://github.com/Leo-emp/operational-efficiency.git
cd operational-efficiency
pip install -r requirements.txt

Usage

# Step 1: Generate the synthetic dataset (2,000 orders, 14,000 records)
python generate_data.py

# Step 2: Run the full analysis (without AI recommendations)
python main.py

# Step 3: Run with AI recommendations (needs free Gemini API key)
export GEMINI_API_KEY=your-key-here
python main.py --with-ai

# Custom output path
python main.py -o my_report.pdf

Get a free Gemini API key at: aistudio.google.com/apikey

Output

After running, you'll have:

operational-efficiency/
├── data/
│   └── operations_data.csv          # Generated dataset
├── charts/
│   ├── 1_efficiency_scores.png      # Scores per step
│   ├── 2_cycle_vs_wait.png          # Time breakdown
│   ├── 3_error_rework_rates.png     # Quality metrics
│   ├── 4_monthly_trend.png          # Seasonal patterns
│   ├── 5_department_heatmap.png     # Department comparison
│   └── 6_prediction_analysis.png    # ML predictions
└── reports/
    └── operational_efficiency_report.pdf  # Full PDF report

Tech Stack

• Python — Core language
• pandas — Data analysis
• matplotlib / seaborn — Visualizations
• scikit-learn — Machine learning predictions
• Google Gemini API — AI-powered recommendations (free tier)
• fpdf2 — PDF report generation

Scoring Methodology

Each process step is scored on 4 dimensions:

Factor	Weight	Measures
Cycle Time	30%	How fast is the step?
Wait Time	30%	How long do orders wait?
Error Rate	25%	How many errors occur?
Rework Rate	15%	How often is rework needed?

Grades: A (80+), B (65-79), C (50-64), D (35-49), F (below 35)

License

MIT