Tagline
A real-time, crowdsourced traffic monitoring and incident reporting platform for developing cities.
Loading portfolio
Description
The Problem Domain
In developing cities like Kathmandu, traffic congestion is a severe daily issue compounded by a lack of expensive, hardware-based traffic monitoring infrastructure like road sensors or CCTV networks. Commuters face unpredictable travel times, and minor incidents often go unreported, leading to massive delays. Existing apps like Waze are tailored primarily for private cars, alienating the large demographic of motorcycle and public transit users.
Technical Implementations
Mobile App: Built a native Android application using the Android SDK to optimize battery life and performance. Integrated the Mapbox SDK to visualize live traffic heatmaps and road incidents.
Web Dashboard: Developed a dynamic Admin Dashboard using React.js for administrators to manage users, verify reports, and monitor system health.
Backend API: Engineered a robust backend using the Laravel (PHP) framework to handle business logic and route data.
Hybrid Database Architecture: Utilized PostgreSQL (RDBMS) for secure user authentication and profile management, while deploying MongoDB (NoSQL) to rapidly process and store high-frequency, real-time continuous location data.
Technical Hurdles Overcome
Data Accuracy & Noise Reduction: Used a "near road probability" algorithm to filter out inaccurate GPS pings and off-road data.
Resource Optimization: Engineered the mobile app to sample and transmit background GPS data continuously without heavily draining the user's battery or consuming excessive mobile data.
Data Integrity: Combated malicious or incorrect crowdsourced data by implementing a fake incident reporting filter and a user-reputation reward system.
Potential Uses & Impact
This system serves as a highly scalable, low-cost alternative to traditional traffic monitoring. It can be used by everyday commuters to navigate peak-hour traffic, by city planners to analyze congestion hotspots, and by local authorities to respond to road incidents faster.
Project story
College final year project
Explaining the idea before building it 🧭
Before I could even start building the Traffic Monitoring System, the hardest part was explaining the problem clearly enough for my supervisor and teachers to believe the project was worth doing.
The product challenge
The idea was not just another map screen. I wanted to show how smartphone GPS samples could become useful traffic signals, and that meant explaining congestion, probe vehicles, location accuracy, speed calculation, and public reporting before the system existed.
The planning work
I prepared a full SRS document, UML diagrams, and implementation plan so the project could be reviewed as a real system rather than only a student demo. The project received an A grade at the end.
The stack decisions
I chose Laravel because its ORM made it easier to keep the backend flexible if the database changed. I chose native Android because Mapbox support and community examples were stronger there than on hybrid platforms at the time, and I chose Mapbox because it fit the student-project budget better than Google Maps API.
A-grade final year project with SRS, UML, Laravel, native Android, and Mapbox tradeoffs.
主なポイント
- 緯度、経度、タイムスタンプ、端末識別子から車両速度を計算しました。
- モバイルとWebの両方でMapboxを使い、交通流レベルを可視化しました。
- インシデント報告、ゲーミフィケーション、公共車両追跡のワークフローを設計しました。
- UML図で挙動を整理し、重要なロジックをユニットテストで検証しました。
担当範囲
- GPSデータモデルとサーバー側の速度割り当てフローを設計しました。
- Android端末をプローブ車両として使うデータ収集コンセプトを構築しました。
- 渋滞とインシデントを確認するWebダッシュボード体験を実装しました。
- システム全体の図面、テスト、UIドキュメントを準備しました。