Each card is someone else’s public dataset, or this station’s own log, drawn on the same dial. The place the data lives is on the card. The questions stay apart, so the page never folds them into one number.
The public repo is RogerWillko/impersonator-weather. Its description is “A public weather gauge for accounts that score like Elon impersonators. Read only.” The default branch is main. The log is counts.json at the root of that branch.
The gauge, this page, and the input file live in the site folder. On this computer the station card reads counts.json beside that folder. The public page reads the raw log on main.
Where the readings sit
Impersonators
The impersonator weather is a stated range, with this station’s log still empty beside it. On 11 July 2026, Jamazel (@JRAzeltine) wrote that active impersonators are likely in the low thousands, and that dormant fakes run to the tens and hundreds of thousands. On this ruler that is Squall for the active line, from 1,000 up to Storm at 5,000, and Banked through Deep for the dormant line, from 10,000 up to Buried at 1,000,000. This station tried on 28 September 2026 and the public search came back as a login wall, so its card reads No reading. One published note and no second list of handles means there is no shared integer.
Followers
The follower weather is Whiteout on both dated counts. In May 2022, SparkToro and Followerwonk published 70.23% of @elonmusk’s 93,452,093 followers as unlikely to be authentic, active users who see his tweets. In August 2023, Travis Brown’s gist counted 62,500,641 follower IDs with zero tweets. Both are past 20,000 on this dial. Jamazel’s note points at those older audits, about 70% inauthentic or inactive and about 40% with zero tweets, and then states the impersonator ranges above them. A follower card answers who follows @elonmusk. An impersonator card answers who looks like him.
Active ruler: Clear 0, Breeze 1, Wind 200, Squall 1,000, Storm 5,000, Whiteout 20,000. Dormant ruler: Quiet 0, Banked 10,000, Deep 100,000, Buried 1,000,000. A range needle sits in the middle of the span the words cover. A count past Whiteout pegs at the end of the dial.
Impersonator handles
Accounts that look like Elon. Active on this station means the account was under 90 days old on the day it was logged.
X sent the search page to a login wall. The count was not reset.
This card follows the log. The sentences at the top were written on 28 September 2026, when that log had no number. The specimen on the station gauge is a drawing and is kept off this page.
Active impersonators/scams: likely low thousands at any time. But dormant/inactive fakes? Tens to hundreds of thousands more.
Key data: Older audits showed ~70% of Elon’s followers fake/spam/inactive. Recent analyses: 40%+ of his followers have 0 tweets or near-zero activity.
Squall
1,000–4,999
active · low thousands
His words, mapped onto the active ruler
Banked–Deep
10,000–999,999
dormant · tens to hundreds of thousands
The same post, on the dormant ruler
The post published words, not a single count. Low thousands is the Squall band. The dormant words cover Banked and Deep. The needle sits midway along that span. Read on 28 September 2026. This card does not refresh.
Followers of @elonmusk
Accounts that follow him. These counts are three and four years older than the station, and they answer a different question than the cards above.
Followers counted 14 May 2022 · Casey Henry, Rand Fishkin, Marc Mims
Whiteout
65,631,405
followers · this page’s product
70.23% of 93,452,093
The blog printed 70.23% of 93,452,093 followers (as of 14 May 2022) as unlikely to be authentic, active users who see his tweets. 65,631,405 is this page’s product, that percent times that base, rounded half up.
Of the 26,878,729 followers who had tweeted in the previous 90 days, the same post published 23.42% as likely fake or spam. This page’s product there is 6,294,998. That product is also past Whiteout. It is still a count of followers.
The same post’s sample of the whole service was 8,555 of 44,058 recently active public accounts, published as 19.42%. That figure is context on this card. Their definition: accounts that do not regularly have a human personally composing tweets, consuming the timeline, or engaging. Read on 28 September 2026. This card does not refresh.
Scrape finished 17 August 2023 · oldest snapshots 27 July 2023
Whiteout
62,500,641
followers with zero tweets
40.8% of the accounts still active in the snapshot
The gist lists 153,209,283 follower IDs. The table gives 62,500,641 with zero tweets, and he states that as 40.8%. The same table gives 65,312,713 with zero followers, stated as 42.6%. He also counts 38,035,372 using the default profile image, which he describes as just under a quarter.
These are published counts, not a product of this page. Matt Binder’s Mashable article discusses the gist. The dataset on this card is the gist. Read on 28 September 2026. This card does not refresh.