FakeOrNot
Crowd-powered content verification system
What It Does
FakeOrNot is a community-driven platform where:
- Users vote on content authenticity
- Admins verify controversial cases
- Trusted users gain voting weight over time
How It Works
1. User Votes
Community members classify content as Real/Fake
2. Admin Verifies
Disputed cases get manual review
3. Earn Points
Accurate voters gain influence
4. Weighted Votes
High-score users impact results more
What will be the algorithm used for point system of users? This.
(I dont have any idea about this algorithm btw)
Technical Stack
Backend
- Go
- MongoDB (user points storage)
Monitoring
- Prometheus (track vote throughput)
- Grafana (admin dashboard)
Anti-Cheat
- One Vote per User per Article
- Admin override tools
System Stats
Cooking
Active voters
Cooking
Admin-verified accuracy
Cooking
Avg. vote processing
Roadmap & Future
- LeaderBoards: Authors with highest number of fakevotes and Users with highest number of points.
- Aggregation Pipelines: Sorting of articles based on recent spikes of votes (trending), most voted etc.
- Algorithm for point system: I dont think that social network algo is relevant here, I dont know any idea about this NOW. Will do it in future, as of now 1 point for everybody lol
- If needed: Will implement my log management tool for batching request and concurrency
Key Challenges
-
Threaded Comments: Tree based structure with
combination of parent-child relationships.
Solution: Use a recursive database schema or materialized path pattern to efficiently fetch and display nested comments. -
Optimised DB Calls: Bulk article payload delivery
over one monolithic request can/will cause high latency, increased
server load.
Solution: Implement pagination and lazy loading to break down large payloads, and use caching to reduce database hits. (but abhi articles hee add nhi huye itne lol)