I noted some variability in the number of how many top items are used in a table of bibliometric analysis paper. Some authors would use top 10, others would use top 15 and even top 20 when they present the most prolific authors, cited papers and journals, institutions etc. I feel it is unlikely to be arbitrary choice in such quantitative discipline! Is it based on the actual volume of literature available? Is there a standard or simply left to open to cover of the most influential publications/authors/institutions ?
Selecting the top 10 among research dataset of 20 is a completely different strategy than selecting the top 10 among research dataset of 2000. So the question cannot be answered this way. Rather, you should consider approach basing on 9th decile, 95th percentile, etc..
Deciding on a list of 'top' items according to some indicator is completely subjective. How important is it, for the reader, to know the top items in this list? If it is of some importance than opting for percentiles, as suggested by Anna Małgorzata Kamińska is a good suggestion. Otherwise you may show a top ten (or twenty or ... ) list, by way of illustration. I think that the best solution is to include a complete list as supplementary material or include the ranking in the dataset as submitted in a data repository.
If you want to be completely objective, look into "Importance Index" theory. It provides an objectively methodology for ranking and applies to a broad variety of topics.
- Badges
- Science topic
- Similar topics
- Papers