Fleet Operations Intelligence Platform

A real-time operational intelligence system for vehicle rental fleet management, featuring anomaly detection, predictive analytics, and business intelligence capabilities.

Live Demo: https://fleet-intelligence.onrender.com/

Overview

This platform provides enterprise-grade fleet operations monitoring with automated anomaly detection, revenue optimization insights, and operational event tracking. Built with PostgreSQL and Flask, it demonstrates advanced database design, complex analytical queries, and data-driven decision support.

Key Features

Operational Intelligence

• Real-time anomaly detection for fraud patterns, maintenance issues, and demand surges
• Automated alerting with severity classification (critical, high, medium, low)
• Predictive maintenance prioritization based on vehicle idle time and rental patterns

Business Analytics

• Fraud detection matrix identifying suspicious customer activity patterns
• Revenue optimization analysis showing pricing discrepancies and profit opportunities
• Fleet utilization metrics with availability tracking and performance forecasting

Data Architecture

• 9-table normalized schema with proper foreign key relationships
• JSON event details for flexible anomaly metadata storage
• Temporal analysis using dynamic date intervals for time-series queries

Database Schema

regions → stores → salespersons
                 ↓
categories → vehicles → transactions ← customers
           ↓
operational_events
performance_metrics

Core Tables:

• regions - Geographic organizational hierarchy
• stores - Location-specific operations data
• vehicles - Fleet inventory with real-time availability
• customers - Client profiles with transaction history
• transactions - Rental records with status tracking
• operational_events - Anomaly detection log with JSON metadata
• performance_metrics - Daily KPIs per store location

Technical Implementation

Backend Stack

• Flask 3.0 - Web framework with Jinja2 templating
• PostgreSQL 14+ - Relational database with JSON support
• psycopg2 - Database adapter with RealDictCursor for row dictionaries

Analytical Queries

Fraud Detection:

SELECT c.customer_id, c.name,
       COUNT(t.transaction_id) as rental_count,
       AVG(t.total_amount) as avg_transaction
FROM customers c
JOIN transactions t ON c.customer_id = t.customer_id
WHERE t.rental_date >= CURRENT_DATE - INTERVAL '30 days'
GROUP BY c.customer_id
HAVING COUNT(t.transaction_id) > 2 OR AVG(t.total_amount) > 500

Revenue Optimization:

SELECT v.model_name, v.price_per_day,
       AVG(t.total_amount / EXTRACT(day FROM t.return_date - t.rental_date))
           as actual_daily_rate
FROM vehicles v
JOIN transactions t ON v.vehicle_id = t.vehicle_id
WHERE t.status = 'completed'
GROUP BY v.vehicle_id
HAVING AVG(actual_daily_rate) > v.price_per_day * 1.1

UI/UX Design

• Dark theme with muted professional color palette
• Responsive layout using Bootstrap 5
• Data-dense interface optimized for operational monitoring
• Real-time metrics with visual severity indicators

Setup Instructions

Prerequisites

• Python 3.8+
• PostgreSQL 14+
• pip package manager

Installation

1. Clone the repository

git clone <repository-url>
cd vehicle-rental-ops

2. Create virtual environment

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

3. Install dependencies

pip install -r requirements.txt

4. Configure database

# Create PostgreSQL database
createdb vehicle_rental_db

# Initialize schema
psql -d vehicle_rental_db -f sql/schema.sql

# Load seed data
psql -d vehicle_rental_db -f sql/seed_data.sql

5. Update configuration Edit config.py to match your database credentials:

DATABASE = {
    'dbname': 'vehicle_rental_db',
    'user': 'your_username',
    'host': 'localhost',
    'port': 5432
}

6. Run the application

python app.py

Access the platform at http://localhost:5000

Or view the live deployment: https://fleet-intelligence.onrender.com/

Application Pages

• Command Center (/) - Real-time KPIs, critical alerts, fleet status overview
• Intelligence Analytics (/analytics) - Fraud detection, revenue optimization, maintenance priorities
• Fleet Data (/browse) - Vehicle inventory with utilization metrics
• Threat Monitor (/alerts) - Operational events log with severity filtering

Project Structure

vehicle-rental-ops/
├── app.py                 # Flask application with route handlers
├── config.py              # Database configuration
├── requirements.txt       # Python dependencies
├── sql/
│   ├── schema.sql        # Database table definitions
│   └── seed_data.sql     # Sample data with anomaly patterns
├── templates/
│   ├── base.html         # Base template with navigation
│   ├── dashboard.html    # Command center page
│   ├── analytics.html    # Intelligence analytics page
│   ├── browse.html       # Fleet data page
│   └── alerts.html       # Threat monitor page
└── static/
    └── style.css         # Custom CSS (if applicable)

Data Patterns

The seed data includes intentional patterns for demonstrating analytics capabilities:

• Suspicious customers with abnormally high transaction values
• Underutilized vehicles requiring maintenance attention
• Pricing discrepancies showing revenue optimization opportunities
• Demand surges triggering capacity alerts