🚲 Real-Time Bike Data Processing

This side project is a technical playground for practicing a Go-based stack with MongoDB, GraphQL. It processes open data on bike availability in Paris, aggregating the number of available docks and bikes over time.

🛠️ Tech Stack

• Go – Backend services and data processing
• MongoDB – Stores raw and aggregated bike station data
• GraphQL – API layer for querying station data

⚙️ Process Overview

1. Fetch real-time bike station data at regular intervals and store it in MongoDB.
2. Use MongoDB Change Streams (watcher) to detect and record station update events.
3. Aggregate updates to compute rolling sums over 15, 30, and 60 minutes.
4. Expose aggregated data through a GraphQL API.

🔗 Data Flow

1. The Collector service fetches data and writes it to MongoDB.
2. The Watcher listens to MongoDB change streams (oplog) and logs events to a dedicated collection.
3. The TimeSeries service aggregates those events into time-based summaries.
4. The GraphQL API exposes both raw and aggregated data to clients.