Each human review is based on Twitter rules that define spam and platform manipulation, and uses both public and private data (eg, IP address, phone number, geolocation, client/browser signatures, what the account does when itβs activeβ¦) to make a determination on each account.