NYC Open Data portal lets users sort datasets by most viewed, but not by how often they are actually fetched through the API. I think it is a better signal since we can actually see which datasets people really use.
Prompt: Don't over-engineer it, Keep it honest. Add a "Most fetched" option to the "Sort by" menu. When selected, the catalog should re-sort datasets based on how often they are pulled via API. The look should stay consistent with the current portal. Please don't lie, and let me know if real API call counts are not available. If that is the case, use the closest usage number you can find and label it clearly.