GamePulse aggregates articles from major games-media outlets via their
public RSS feeds, then uses Aerospike Graph to surface consensus,
momentum, and sentiment signals that basic keyword counting misses.
How it works
A background scraper fetches RSS feeds every 30 minutes from 12+
outlets — IGN, PC Gamer, Eurogamer, Polygon, Kotaku, and more.
Articles are stored locally (time-window retention) and their text is analysed to extract
entities — game titles, companies, platforms, genres, and events.
These entities and their connections are written into an Aerospike
Graph database for cross-outlet trend analysis.
Deep Trends (Graph-Powered)
- Trending Now: Distinctive entities gaining momentum in recent coverage.
- Cross-Source Signals: Same quality filters as Trending Now, plus confirmation across 3+ independent outlets.
- Emerging Trends: Topics accelerating in the last 24h compared with the prior week.
- Trending Sentiment: Coverage tone with mixed-signal handling and gameplay/business disambiguation.
- Source Coverage: Heatmap of which outlets are covering which topics.
- Social Pulse: Reddit + Bluesky trend and sentiment signals using the same model as Deep Trends.
- Media vs Social: Side-by-side media vs social differences in topic share and sentiment.
Tech stack
- Backend: FastAPI (Python) with async RSS fetching + entity extraction
- Graph: Aerospike Graph Service 3.1.1 (Gremlin over WebSocket)
- Database: Aerospike Enterprise 8.1 (graph storage backend)
- Frontend: Vanilla HTML, CSS, and JavaScript
- Storage: JSON file (article cache) + Aerospike Graph (deep analysis)
- Serving: Nginx (static files) + reverse proxy to FastAPI
- Deployment: Docker Compose (4 services)
Privacy
GamePulse has no user accounts, no cookies, no analytics, and no ads.
All article data comes from publicly available RSS feeds.