54.9m Qanon tweet numeric IDs
收藏DataCite Commons2020-09-14 更新2024-07-28 收录
下载链接:
https://figshare.com/articles/dataset/54_9m_Qanon_tweet_numeric_IDs/12952190/1
下载链接
链接失效反馈官方服务:
资源简介:
We began capturing Qanon related tweets as a stream on 21 July 2019. Every time we noticed a new hashtag in use it was added to the list, which currently has 91 terms in it.<br>The monolithic file was enormous, so this set was split into ten pieces by using the last digit of the numeric tweet ID. You will notice that tweets with numeric IDs ending in an even number are twice as prevalent as those ending with an odd number.<br>This odd/even bias is an artifact of how Twitter generates unique IDs for tweets and users. The last twelve bits of every ID are produced by a pseudorandom number generator. We suspect this "snowflake number service" was coded in assembly language and that it employs a barrel shift register. This would be a fatal flaw for an encryption algorithm, but since it's just being used to produce unique identifiers in ascending order, Twitter chooses speed over randomness.<br>Each tweet ID in the csvs is followed by the numeric month, the date, and the year. These "dehydrated" numeric IDs may be restored to full tweets using DocNow's hydrator.<br>https://github.com/DocNow/hydrator<br><br>
提供机构:
figshare
创建时间:
2020-09-14



