← Back to Portfolio
Mobile Ads Car Analytics
Route cost intelligence and real-time analytics platform for mobile advertising vehicles.
Inspiration
Years ago during a traffic jam, I saw a mobile advertising car stuck on the road. That moment sparked a question:
what happens to advertising cost efficiency when the vehicle never reaches its destination on time?
If delays increase operational cost, businesses need a way to predict, analyze, and optimize route expenses before deployment.
This project was built to solve that problem for mobile advertising fleets.
What It Does
- Analyzes cost across routes
- Calculates cost per meter
- Compares cost between vehicles
- Evaluates cost per driver
- Breaks down cost per waypoint
- Detects anomalies in trip spending
How It Was Built
- Real-time analysis agent powered by Google ADK
- Vehicle monitoring system capturing video and route telemetry
- Trip data stored in Firestore
- Analytics engine processes historical + live movement data
Key Analytical Capabilities
Driver Cost Comparison
The system can compare route cost between multiple vehicles with different pricing models.
- Car 1 → $0.22 per 10 km
- Car 2 → $0.30 per 20 km
Returns:
- Total cost per route
- Waypoint breakdown tables
- Cost difference analysis
- Most expensive vehicle identification
Trip Intelligence
- Total trips per vehicle
- Driver time-on-road comparison
- Latest trip analytics
Fuel Cost Analysis
Supports multiple pricing models:
- Cost per km
- Cost per meter
- Variable rate per vehicle
Outputs:
- Leg-level cost tables
- Total fuel cost
- Graph coordinate output for visualization
- Anomaly detection datasets
Prompt Examples
Calculate total driver cost for Car 1 if rate is $0.22 per 10 km.
Return waypoint table with:
distance | from | to | cost
Compare with Car 2 if rate is $0.30 per 20 km.
Return difference and identify most expensive vehicle.
Calculate fuel cost for latest trip of Car 1
Fuel price = $3 per 18 km
Return table + total cost
Generate graph coordinates for Car 2
x = cumulative distance
y = cost per leg (not cumulative)
Challenges
- Cloud infrastructure cost optimization
- Limited documentation for new frameworks
- Real-time data synchronization issues
- Telemetry reliability from moving vehicles
Accomplishments
Built a fully working analytics engine capable of predicting route costs, analyzing driver efficiency,
and comparing operational expenses across vehicles in real time.
Technology Used
- Google Agentic ADK
- Google Gemini
- Google Firestore DB
- Google Cloud Run
- Google Maps API
- ReactJS
- Node.js
- Vite
- Google Veo
- Python
- Gradle UI
- Google Cloud Storage
- Google Service Account
- Google Artifact
- Docker
- Microservices