<
>
If were considering data for a whole year, the top range, 10 posts (although +) is less than an average of 1 post per month on the lower end of the range, which is a small contribution, and perhaps gets too mixed-up in the same segment with the real heavy posters.
Finding the adequate groups is not trivial. In my case, since Im only focused on one single board, Its easier to create groups that are more fitting to be meaningful. With an overall approach, Id probably have created the group based on conceptual average weekly posting ranges, and scales that up to a yearly outcome.
i.e. Tentative:
+10 posts per week -> [500++] posts per year (Heavy Regular posters)
[5..10) posts per week -> [250..500) posts per year
[1..5) posts per week -> [50..250) posts per year
(0..1) posts per week -> [1..50) posts per year
Now after trying a few local boards, it may result in the bottom segment being too concentrated (i.e. the Spanish local board would render 94,62%), and Ive got a feeling that there may be many 1..2 posts per year accounts in general. It may be worthwhile breaking down the lowest group to see this hypothesis (which is meaningful per se), as it turned out to be in the Spanish local board:
(0..1) posts per week -> [1..50) posts per year could be replaced by:
[26,50) posts per year
[3,25] posts per year
[1.2] posts per year
As I said, the above is tentative, and subject to what you see in the data, and how much of an effort it implies. Perhaps just to emphasize that its interesting to separate the heavy posters and the very sporadic posters as separate groups (as I did above), and break down all between figures into something manageable.