As per Willie Wong's comment I got these results for Tags activity over time in table format.
- For the (divergent-series) tag $\quad$
7, "2011", "5"; 4, "2011", "6"; 4, "2011", "8"; 1, "2011", "9"; 7, "2011", "10"; 12, "2011", "11"; 2, "2011", "12"; 4, "2012", "1"; 4, "2012", "2"; 1, "2012", "3".
- For the (real-analysis) tag $\quad$ 7,"2010","7"; 47,"2010","8"; 35,"2010","9"; 54,"2010","10"; 86,"2010","11"; 59,"2010","12"; 77,"2011","1"; 51,"2011","2"; 110,"2011","3"; 91,"2011","4"; 87,"2011","5"; 106,"2011","6"; 78,"2011","7"; 110,"2011","8"; 144,"2011","9"; 213,"2011","10"; 249,"2011","11"; 185,"2011","12"; 208,"2012","1"; 271,"2012","2"; 153,"2012","3".
The details do not include the time, as far as I understand. However it is a part of what I had in mind when I asked the question.
In the meantime Willie Wong posted another idea to be explored.
Added. The "Tag frequency by Time of Day and Day of Week" is obtained as per Willie Wong's answer, which seems enough for the purpose of my question. See his answer.
Edit to add. For the (calculus) tag I ran a query in "Tag frequency by Time of Day and Day of Week". The plot is this one.
