Home > Web/Tech > Some Fun Web Data and Some Light Duty Analysis

Some Fun Web Data and Some Light Duty Analysis

Some quick notes on Google data for the University of North Dakota Website.  I did this mostly for fun, but the results are sort of interesting…

Google Trends

The goal of this short report is to summarize my very superficial analysis of our site data via the Google Trends interface. I suspect that this report will raise more questions than answers, but some of the patterns in the data are interesting (and perhaps vaguely alarming). As with all things Google, they are relatively opaque concerning what variable go into their analysis, but I suspect that they are more or less stable and the analysis is systematic.

Analysis of und.edu

This first graph shows the number of visitors who visit the website over a 1 day period. This chart compares und.edu (in blue) with ndsu.edu (on red), and as a control grandforksherald.com (in gold). The most curious thing about these graphs is the huge drop in traffic over summer of 2008. This could be a simple matter of the data that Google for analysis, but if it is associated with a particular change in the way that the site was organized, the we should make note that visitors stopped visiting our main domain (i.e. und.edu) rather abruptly and most never returned.


Moreover, the general trend, according to page rank, is that our site (as well as ndsu’s site) has become less popular through time. It is interesting to note, however, that the downward trend appears to be the case will any number of university home pages. This chart adds sdstate.edu (gold) and wichita.edu (green) to the chart. The similar slide over the summer months of 2008 suggest that the pronounced slide in und.edu has at least something to do with how Google collected data.


To compare the decline in the number of visits to main domains, here is a chart comparing four major big ten universities: osu.edu (blue), psu.edu (red), umich.edu (gold or, better, maize), and umn.edu:


They all show similar declines with a slide during the summer months of 2008 (although this slide is far less pronounced at osu.edu).

The rather steady decline of the und.edu domain also appears in the data prepared for Google Ad Planner. Of course, in this context Google is trying to sell us on advertisements, so they have every interest in showing declining visitors, but it is nevertheless interesting:


To this can be added some basic demographic data collected by Google. I take this cum grano salis, but it is interesting to contemplate and there might be so real motive for Google to be accurate here. They make money (actually, almost all their money) from per-click advertisements. So the more people who click on your advertisement, the better they do. Consequently it is in their best interest to provide the user with good data to maximize the visibility and profitability of their advertisements.







While the data from these charts is probably inconclusive, it seems to suggest that visitors to the main domain of university pages are on the decline. This may well reflect the proliferation of servers on campus (and multiple domains), but I suspect that it also reflects changes in how the web is surfed, with visitors less frequently jumping from main page to main page and more frequently entering into domains through numerous other entry points.

Google Insights for Search (beta) Data

This service provides data on “the number of searches for a particular term relative to the total number of searches done on Google over time” normalized on a scale of 0-100. To begin, I compared the search terms “University of North Dakota” (light blue) to “North Dakota State University” (red). Since these are specific search terms, it is probably worth noting that NDSU is probably better known was “North Dakota State” than “North Dakota State University”, I included searches for “North Dakota State” (gold). I would guess that the higher “North Dakota State” numbers are bolstered by people searching for things other than the NDSU, like state offices or even general information on the state itself.


This graph shows the relationship between the overall number of searches in the specific category of colleges (dark blue) and universities to searches for “University of North Dakota” (light blue), “North Dakota State” (gold) and “North Dakota State University” (red); the dark blue line represents the relative growth of the number of specific searches against searches for colleges and universities in general. Apparently “University of North Dakota” performs better as a “global brand” than in the subset of university and college searches. Note the huge spike in searches for North Dakota State and North Dakota State University precipitated by the flooding last spring!


It is important to note that the downward sloping
line does not necessarily mean that the number of searches for “University of North Dakota” is declining, but that the number of searches for this term in relation to the number of searches through Google overall is declining. The line however does decline at a significantly steeper slope than the lines for either “North Dakota State” or “North Dakota State University”. This suggests that the number of searches for our university name have declined more significantly in relation to all searches than searches for our friendly rival to the south.

As a check on this, I also compared Google searches for UND versus those for NDSU, recognizing, of course, that und is a very common word where I don’t thing ndsu means anything (it may be a kind of Asian fungus, but even then a rather rare one). To compensate for this I limited our comparison to the U.S. where people use the word und somewhat less frequently than, say, in Germany. I also did a comparison for North Dakota and Minnesota. UND is light blue and NDSU is red.

In the US:


In North Dakota (with forecasted data):


In Minnesota (with forecasted data):


The results here were surprising. Despite the fact that “UND” is a word, it is still being searched for less than “NDSU” in North Dakota and they are very close in Minnesota.

As another point of comparison for the analysis of various searches, I produced a chart based on searches for “Penn State” (blue), “Ohio State” (red), “University of Michigan” (maize) and “University of Minnesota” (green). The very dark blue line shows the overall searches for the Colleges and University Category.


It clear that the searches for the name of the university reflect the rhythm of academic life (and presumably recruitment), with peaks in September and October. To see whether I could isolate other, non-academic influences, I plotted searches for WCHA (red), “Fighting Sioux” (gold) and “University of North Dakota” (blue). To be safe, I checked the WCHA against searches for “Frozen Four” and found them almost completely parallel. I also checked “Fighting Sioux” against “Sioux Hockey” and found that they carved a similar pattern. In any event, it is interesting to note how little the Fighting Sioux’s end of the season battles has on searches for the term “University of North Dakota”.


Like the charts based on visits to various specific pages and domains, these charts (which show specific searches) also show a decline of searches for broad forms of classification. This chart reflects the decline in specific searches for the name of the university (i.e. University of North Dakota) and it more or less coincides with the decline in site visits to und.edu as a specific domain. The best analysis of this correlation is that visitors to our site (which according to our in-house Google Analytics data have remained relatively stable) are entering through venues other than the main und.edu page and they are not searching for the university through general search terms like “University of North Dakota”.

Categories: Web/Tech
  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: