“…The error grows considering more specific hours (e.g., [06,11] and [12,15]), during which the number of submitted check-ins is relatively small (see Table I). Indeed, in those cases, it is more difficult to recognize the top-k venues, since the introduced noise tends 11] 12,521 [12,15] 26,127 [16,19] 56,468 [20,24] 65,557 [00,24] 231,229 to be of the same order of magnitude of the true value (number of check-ins submitted during those hours), thus reducing the utility of the data. However, when considering time-specific queries, the average error is well below 10% using values of j ≤ 2.…”