My very limited time on X has already shown that posts ranked by number of expression is highly non-linear. Maybe Zipf or Pareto distributed? The first plot shows each post sorted by impressions (rank 1=most impressions). You'll see a steep drop from ...