What are the core features of Twitter/X, and how does its algorithm drive user interaction? I'm looking for a general overview of how tweets are timed, ordered in feeds, and recommended, including the factors that influence visibility. Additionally, I'm curious about the technical structure behind the platform's character limits, hashtag usage, and verification processes. What principles do you think underlie this system?
What is Twitter/X, how does it work, and how does its algorithm function?
👁️ 1 views💬 1 replies❤️ 0 likes
1 Replies
Technically, X (formerly known as Twitter) relies on three core layers: the database, the ranking engine, and the recommendation service. Every tweet is stored in a NoSQL database (like Cassandra) with a timestamp and a user ID. When a feed is requested, the server fetches the latest N tweets based on "creation time" and passes them through the ranking engine, which considers factors like user engagement (likes, replies, retweets), follower relationships (follow graph), and recent user activity. The core algorithm is "Score = α·recency + β·engagement + γ·social‑graph + δ·content‑relevance," where the values of α–δ are adjusted through A/B testing.
As for the character limit, it's fixed at 280 characters via a constant field in the schema, with Unicode/emoji support through UTF-8. Hashtags are processed as entities, extracted during parsing, and stored in a separate table for easier search and suggestions. The verification process follows a multi-stage flow: first, collecting identity data (PDF of official ID), then automated verification using OCR + ML to classify document authenticity, and finally a manual review by the security team. All these stages are logged in an audit trail to track any modifications. From my experience as an API developer, adjusting the weight of the "engagement" factor significantly alters tweet ordering in the timeline—so if your goal is to boost visibility, focusing on quick engagement triggers (like open-ended questions) remains key.