Statistics gathered daily could contain both daily cycles and weekly trends. This can lead to chaotic-looking graphs when activity is plotted day by day. You can improve your graphs easily using SQL.
Look at Figure 4-1, which shows a chart of the raw figures for the number of page views for a web site per day, over the course of one year. These figures come from Webalizer, the web log analysis program (http://www.mrunix.net/webalizer). It is difficult to see the trends because the weekly cycle overwhelms the daily detail, and obscures the long-term trend.
Figure 4-1.Page views per day over one year
To understand the data, you need to separate the effect of the weekly cycle from the table. You can see the weekly cycle by taking the average for Monday, the average for Tuesday, and so forth. In Figure 4-2, Monday to Sunday are numbered 0 to 6.
Figure 4-2. Average page views by day of week
Notice that the value of the Sunday column (column 6) is less than half the value of the midweek columns. This is helping to cause the zigzag pattern in the original graph. If you view the data averaged per week (see Figure 4-3) rather than per day, it is easier to see the long-term trend.
Figure 4-3. Smoothed data--page views averaged for one week
The graphs and charts shown here come from Excel. Many spreadsheet applications, including Excel, have tools for importing directly from databases and producing a variety of graphical reports.