The first statistics shows the number of accesses to files of certain subject areas of my homepage per week between summer 2000 and april 2003. The interest in ticks (Zecken) certainly is seasonal. It is highest im early summer. A similar thing hold for the magpies (Elstern), which are most active in april and attract the attention of garden owners. The visitors of the moss-page (Moose) are probably mainly pupils and students. Hence the minima during the summer and chrismas.
Compared to those the curves for the recipes (Rezepte) and the computer stuff are more even. The increase of the accesses to cooking recipes, which nearly didn't change over the time, should show more or less the increase of the internet use generally.
This statistics shows the distribution of accesses to files of different format over the day (broken down to hours). E.g. one can see, that every html file in average loads 1.5 jpg-files. But in the night the "background-noise" of the indexing robots, which normally only request html files, dominates.
Here I choose five arbitrary search engines or web-directories and counted visitors coming from these. The 100% don't refer to the total number of visitors, but only to the number of visitors coming from one of those 4 engines. Furthermore those 4 are not disjoint, e.g. accesses coming from google.yahoo are counted twice.
I don't know the reason for the nearly absent accesses from lycos in the year 2000. Perhaps its indexing robot simply needed somewhat longer. Or perhaps the reason is, that the statistics for 2000 isn't really meaningful, because the quantity of data isn't sufficient.
To get out the used browsers from the log files is possible up to a certain degree, but remains a rather uncertain thing. E.g. the different versions of the Internet Explorer identify really nicely, but there are other browsers, which pretend to be Internet Explorers, because there are servers, which refuse other browsers than IE and netscape. But in total the statistics should agree more or less with reality.
The get out operating systems from log files is even more difficult than to get out the browsers, and in most cases not possible. At least the rather common Windows-95, Windows-98 and Windows-NT systems are easy to recognize. Here is a statistics which shows the percentage of accesses from those three distributed over a day (broken down to hours). It can be assumed, that many NT-users access the page from their work, whereas 98-users rather surf from their home.
In the course of the qualification for the european football championship a match germany-faroes, which germany won 2:1, took place on wednesday 16.10.2002. This pictures shows the number of accesses per hour of my two faroes-sites ( http://www.ijon.de/foroyar/ and http://www.lamedon.de/urlaub/faeroeer/) between tuesday and friday of the week. The vertical blue line marks the moment of the starting whistle.
The two faroes-sites made about the half of all accesses in this week.
Googles program, which searches the web sites and puts them into the index, identifies itself very clearly in the log files as Googlebot/... As can be seen here, it scours my site relatively regularly every 4 weeks. The reason for the outstanding peak in april/may 2002 is unclear.