8.11 Example: range analysis of web hits

Suppose Acme has a web site and wants to view information on who and how often people visit their web pages. Also, suppose Acme has saved all of its log files for the past three years.

The Acme system administrator could write a quick Qtcl (i.e. Qddb tcl script using Fx interface) script that parses the web log files and populates a Qddb database. Let's assume the system administrator created a schema shown in figure 8.9.

Use cached hashing
CacheSize=100
HashSize=20000
Use reduced attribute identifiers
Use Cached Secondary Search

Domain separators ""
Hosts (
    Name separators ""
    Accesses (
        Date type date verbosename "Date/Time" 
            format "%m/%d/%y %H:%M:%S"
            Page separators " 	/"
    )*
    Number type integer verbosename "Number accesses (host)"
)*
Number type integer verbosename "Number accesses (domain)"
Comments

  
Figure 8.9: Web hit schema

This schema allows you to look at (1) total number of hits, (2) number of hits by host and (3) number of hits by domain.

Let's assume you already wrote your Qtcl script and now have a Qddb database populated with all the hits over the past three years.

By now, you should be familiar with generating reports. If you need more practice with generating reports, refer to chapter 7.

For all of the graphs and barcharts in this section, choose the following settings:

Range analysis checkbutton

should be turned on.

X-axis attribute

should be set to Hosts.Accesses.Date

Y-axis attribute

should be set to Hosts.Number

Now, for each different graph we'll just change the Range/Format parameters by clicking a different pull-down option. The following sections illustrate a few of the ways to view your web site data hits.

8.11.1 Graph of daily hits by month

Choose the Month of year (3-letter) option to display the graph shown in figure 8.10.

  
Figure 8.13: Generated from ``Month of year (3-letter)'' option

8.11.2 Graph of average daily hits by day of week

To obtain the graph shown in figure 8.11, choose the ``avg'' value from the range operations pull-down menu and then choose the Day of week (3-letter) option.

  
Figure 8.11: Generated from ``Day of week (3-letter)'' option

8.11.3 Barchart of yearly summaries

Choose the ``sum'' value from the range operations pull-down menu. Then, choose the ``Year'' pull-down menu option. After applying these changes, you'll see a chart similar to one shown in figure 8.12.

  
Figure 8.12: Generated from ``Year'' pull-down menu option

8.11.4 Barchart of summary by month

Be sure the ``sum'' value is chosen in the operations pull-down menu and then choose ``Month of year (3-letter)''. After applying these settings, your barchart should resemble the one shown in figure 8.13.

  
Figure: Generated from ``Month of year (3-letter)'' option

8.11.5 Barchart of summary by day of week

Be sure the ``sum'' value is chosen in the operations pull-down menu and then choose ``Day of week (3-letter)''. Your barchart should resemble the one shown in figure 8.14.

  
Figure 8.14: Generated from ``Day of week (3-letter)''

8.11.6 Summary

This section has provided the steps to performing range analysis in creating graphs and barcharts. We have presented a detailed scenario of how to create useful representations of data for a fictitious company. Specifically, we looked at a sample schema for maintaining web site log information and how to generate different views of this data using Qddb's powerful range analysis tool.