HAAGE&PARTNER Computer GmbH  HAAGE&PARTNER

Sawmill Analytics

Analyse und Reporting
für Web | Netzwerk | Sicherheit

Zugriffs- und Datenanalyse von Server-Logs (Proxy, Mailserver, Firewall, Webserver) und Überwachung der Sicherheit & Performance, Schwachstellenanalyse.

Sawmill Analytics 8 | Loganalyse

Sawmill-Tutorial

Using Zoom To Get More Detail


When you first view reports in Sawmill, you will see the Reports Menu along the left of the page, with a list of a few dozen available reports. These are unzoomed reports, showing the "top ten" of each field in the database. For instance, in a media server analysis, you might see the top IP addresses in one report, with bandwidth and viewing time for each, or you might see the top publishing points, with bandwidth and viewing time for each. These are valuable reports, and a good starting point for any investigation, but Sawmill can give a lot more detail than you see in those reports. Because the power and flexibility of "zoom" is not always obvious, this newsletter is dedicated to describing how "zoom" works, and to exploring some of the more advanced zoom options.

For this example, we will use a web server dataset from 1998, for a web site which publishes reviews of novels. Clicking Pages/directories, we see the top-level directories and pages of the site:
page_directories
The Pages/directories Report


The top one, /seized/, is the main directory of the site, we we click on this. Clicking on it zooms, which means it applies a filter to the dataset, and possibly switches to a different report (we'll discuss this report switch more below). In this case, it filters the data to show only hits on the /seized/ directory, or files in it. When zooming on a hierarchical field like the "page" field, the default behavior is to stay in the same report, so Sawmill zooms to /seized/, and redisplays the Pages/directories report with this zoom filter, shown in yellow. The result of this is that it feels a lot like zooming into the folder structure on a hard drive--you click the folder name to see what's in it, and here we've clicked the directory name to see the contents:

pages_directories_zoom_seized
The Pages/directories Report, Zoomed On "/seized/"


In this web site, the reviews themselves are in the /seized/reviews/ directory, so we click that to zoom in another step, to see the contents of /seized/reviews/:

page_directories_zoom_seized_reviews

The Pages/directories Report, Zoomed On "/seized/reviews/"


There are 122 items in this table (only the top ten are shown), and if we looked further, we would see that most of them are review pages (the name of the novel, followed by the name of the reviewer). Rows 8, 9, and 10 here show the most popular reviews (ignoring Practical Magic for the moment), with 120, 126, and 155 page views, and if we looked further, we would see a smooth drop from that, toward the less popular reviews. But way out ahead of the pack is /seized/reviews/practical_magic_sara_lipowitz.html, the review of Practical Magic. This is an anomaly--this review has many times more page views than any other review. Why? Sawmill can help you find the answer. Let's start by zooming on that review, by clicking the second row. Since that's a filename, we can't zoom further in the Pages/directories report, so Sawmill automatically zooms to the Overview instead:

overview_seized_reviews
The Overview, Zoomed On "/seized/reviews/practical_magic_sara_lipowitz.html"


This isn't very interesting by itself (the data here is mostly the same as the row of the Pages/directories table), but it is a good staging point for further investigations. The key to further zooming is the "Zoom to report" menu, which appears below the yellow zoom description. We can select any report from that menu, and it will display that report, while preserving the zoom. This is different from what happens if we click the report in the Reports Menu, because that discards the zoom and goes back to the top-level report. By using the "Zoom to report" menu, we can break down the data on any field, finding out more about this particular subset of the data. Let's start by selecting Days from the "Zoom to report" menu. This shows the Days report, subject to the current filter:

days_zoom
The Days Report, For Practical Magic


This report shows traffic on just that one file, day-by-day. The graph at the top shows that there was a large spike of traffic in mid-October, 1998. Before that, traffic on this novel review was very light; after that, it was much higher. So what happened in mid-October 1998? A little web research shows that was the release date (October 16, 1998) of the movie version of the novel Practical Magic. With the movie's release, the novel got much higher exposure than before, which sparked a sharp interest in the review. If the goal of this web site is the bring the maximum number of page views, then, this gives a clear recommendation for which novels to review: review those which are being made into movies. Sawmill's detailed analysis can give similar information for any web site, information which can be used to make the site more effective, or more popular.

Just for the sake of demonstration, let's do a little more digging. From "Zoom to report", select the "Domain descriptions" report. This shows the domain descriptions where traffic came from, to the review (again, we're still zoomed in on just this one review page, so we're seeing a very specific report: domain descriptions for the hits on Practical Magic):
domain_descriptions
Domain Descriptions For Practical Magic


Much of the traffic was from .net, .com, .edu addresses, and IP addresses. But somewhat surprisingly, there are some hits from Singapore (*.sg hostnames). Let's look deeper, by clicking "Singapore (sg)", and zoom to the Hostnames report:

hostnames

Hostnames from Singapore, For Practical Magic



This shows a list of all hostnames of the browsers who accessed this page from Singapore. Now, let's zoom on milkyway.singnet.com.sg, to see the specific events from that hostname. But this time, we'll save some time by clicking the Zoom Options tab, and selecting "Log detail" below it (this is usually faster because it takes several seconds to generate the Overview, but no time at all to display Zoom Options):

zoom_to_report_log_detail

Hostnames from Singapore, For Practical Magic, With Log Detail Zoom


That indicates that we don't want to zoom to the Overview (the default), and then zoom from there to "Log details"--instead, we want to zoom directly to "Log detail." So now when we click milkyway.singnet.com.sg, we go straight to the "Log detail" report, and see full details of those four page views from milkyway.singnet.com.sg, on /seized/reviews/practical_magic_sara_lipowitz.html, including the exact time of each hit, the referrer, and more (additional fields have been truncated to fit here, but all database fields are in this report):
log_detail

Log Detail For Practical Magic, from milkyway.singnet.com.sg


This type of deep forensic analysis is useful for any type of log data. Any number of zooms can be applied simultaneously, and can be used in conjunction with other types of filters, including date range filters and global filters. Zooming can continue on any number of fields, to any level, including the level of the events themselves, in "Log detail."


Advanced Topic: Changing The Default Zoom For A Report

In the example above, we zoomed by clicking on an item to zoom to Overview, and then selecting a report to zoom to. Later, we saved some time by using the Zoom Options tab in the report. But if we know we'll usually be zooming from Report A to Report B, we can modify Report A so its default zoom is to Report B, rather than to the Overview. This effectively changes the Zoom Options menu selection, so we don't have to do it manually if we just want to zoom to Report B. For instance, we could change the zoom default on the "Domain descriptions" report to zoom to the "Hostnames" report, so any time we click a domain, we'll see a list of hostnames under that domain. This is done by going to the Config page for the profile, then going to Manage Reports, then Reports/Reports Menu, then clicking the report name, clicking the Report Elements tab, clicking Edit to edit the report element, and choosing the destination report from the Default report on zoom menu. So after changing the default report on zoom to Hostname in the "Domain descriptions" report, the report element editor page would look like this:

default_zoom

Changing Default Report On Zoom To "Hostnames"

Save the change, and in the future, any click in the "Domain descriptions" report will zoom to the "Hostnames" report.




Professionelle Dienstleistungen

Sollten Sie die Anpassung von Sawmill Analytics nicht selbst vornehmen wollen, können wir Ihnen dies als Dienstleisung anbieten. Unsere Experten setzen sich gerne mit Ihnen in Verbindung, um die Reports oder sonstige Aspekte von Sawmill Analytics an Ihre Gegebenheiten und Wünsche anzupassen. Kontakt

Zur Tutorial-Übersicht

Weitere Informationen

      Live-Demonstrationen »    
© 1995-2011 HAAGE & PARTNER Computer GmbH · Impressum · Datenschutz · www.haage-partner.de