Methodology

Data sources

We collect public X (Twitter) posts for a fixed set of tracked keywords and accounts per niche. Each post is stored with its engagement counters (likes, reposts, replies, views) and the author's follower count. We never fabricate posts or numbers: every value on every page is read from our own database. Posts may be written in any language; non-English posts are shown in their original text and tagged.

Snapshot cadence

Post counters are re-read every 15 minutes for posts published in the last 48 hours, and each reading is stored as a new snapshot. Author follower counts are recorded once a day for every tracked creator. Snapshots are append-only: history is never overwritten, so velocity and growth are measured from real changes over time.

Heat score (0–100)

The heat score blends reach, engagement, velocity and freshness. Each component is squashed onto a 0–1 log scale so one viral outlier cannot pin the board at 100. Freshness zeroes the whole score for posts older than 48 hours, so “Rising” really means recent.

Bucket labels: Breaking (80–100), Hot (55–79), Warming (30–54), and Rising (below 30).

Velocity and growth

Velocity is new likes or views per hour, measured between the two most recent snapshots. Follower growth is the change in followers across a 7- or 30-day window of daily snapshots. A growth number is only shown once the window actually has history; until then the column reads “Collecting data” rather than showing an inaccurate figure.

Limitations

• · We track a fixed set of keywords and accounts per niche, so this is a sample, not a census of X.
• · View counts are only available for some posts, so velocity can be partial.
• · Follower growth needs several days of snapshots before it is meaningful.
• · X Analytics is independent and not affiliated with X Corp.