The trending tag cloud uses sampling and summarization to improve performance when computing the trending tag cloud.
All activity in the sample interval, is compiled into a single row per tag. The trending window is then the number of minutes into the past to consider if a sample should be included into the trending tag cloud statistics.
An example of how the sampling interval works:
So if you have 5 replies in the first 5 minutes of the hour, and then 10 replies in the second half of the hour, and then 5 replies in the last 5 minutes of the how; those usages are stored as:
- 5 usages in the first half an hour.
- 15 usages in the 2nd half an hour.