Web Analytics

other sections

teaching | scholarship
service | mentoring

other pages this section

scholarship home
discovery: fiction | video
teaching: webs | Ricci Street

this page

how big is Ricci Street? | who uses Ricci Street?
how many visitors? | Parkside Plaza student webs

From 1998 until 2004, AIT gave me 600 MB of server space, since then, 3 gigabytes, and it is always full. I could get more, but I'd just fill that, too. Periodically, I delete student files at Parkside Plaza, after notifying the students, of course, in order to stay under the limit.

How big is Ricci Street?

Thousands of document and image files are stored on the server in the riccistreet folder. Over the years, the student sections have grown the most. A large number of Medaille students, especially MBA's, got their first taste of publishing, of participating in the online conversation, in my courses.

How big is my part?

Because of the fluid, non-standardized nature of information on the web, this table has a snapshot from September 2003 followed by a guesstimate. These are files that I made with FrontPage and FTP'd to the server. They are not files made and published by students and other faculty. Since 2003, additions have outnumbered deletions.

The document files are almost all .htm files. Some of them would print out to one or two 8 1/2" x 11" ink-on-paper sheets. Many of them would print out to more than ten or twenty 8 1/2" x 11" sheets. I estimate that the nearly thousand documents / web pages that I have made and published on Ricci Street and are still there as of September 2003 would print out to over five thousand sheets of paper with a 10 pt type size. If I included everything that I wrote and published over the last decade and that has been deleted, I would have to double that quantity.

As for the links, the numbers above came from FrontPage's reporting function, examining all the files on the copy of Ricci Street that is on my laptop. The internal links are often parts of templates that I made only once. However, the twenty thousand external links still active as of September 2003 is a number that seems too large until you divide it by the four-plus years, or fifty months, that RicciStreet was active until then. Over that time, it is only four hundred links that I made per month, a hundred per week, which is quite reasonable. But they sure add up.

Who uses Ricci Street?

Server Traffic

Every request to a web server is recorded in a log file. Over the years, I have downloaded the logs monthly. In 1999 when I started using Web Trends analytic software, I was fascinated by what I could learn. From then on, I spent at least an hour in every marketing course showing students what online merchants can learn about the behavior of their customers.

After I had analyzed the trends every month for a couple of years, I finally saw enough patterns to figure out some changes that I needed to make in Ricci Street's structure. Some of them, I could make at the time. Other changes, more complex, will need the occasion of a total site re-design to make worthwhile.

Since then, I have checked the logs only a couple of times a year, and then only to look for anomalies and outliers. However, when I do come to re-design Ricci Street or, better yet, abandon it and start over, then a long, thoughtful look at some detailed traffic patterns will be very helpful.

Web site traffic is notoriously easy to manipulate. In ten minutes of fancy coding today, I can double next month's site traffic. However, I am studying human behavior, not trying to impress a client, so I have everything to lose by such manipulation. For purposes of comparisons over time, I have made every effort not to make any arbitrary changes to Ricci Street that would affect the traffic counts one way or the other. I have also used the same software with the same settings to analyze all of server logs that produced the tables below.

For an example of more detailed WebTrends output, see the March 2007 report.

For a discussion of the complexities and trade-offs of server logs, see the Site Stats section of Ricci Street, especially the summary of the first three years.

The table below has some summary numbers. If you would like to see the raw log files, please email me. To note on the table below:

The dips and leveling in 2002/03 is due to moving all the student webs from the Parkside Plaza section of Ricci Street < RicciStreet.net/dwares/plaza/ > to its own domain, ParksidePlaza.net, on a server in Hong Kong. Unfortunately, while the price was right, I did not have access to server logs.

A consistent 15 - 20% of the traffic  (including some of the Unknowns) is from non-US-based IP numbers.

Half a dozen countries turn up in the logs every month, usually in this order:

United Kingdom, Canada, Netherlands, France, Germany, Singapore, Australia, New Zealand

The foreign traffic in November 2005, for example, came from 98 different countries. Over the course of a year, almost two hundred different countries make an appearance in the server logs.

June 2006 shows persistence of non-student traffic. One way of looking at it is, Look at how much traffic drops off when the students stop using Ricci Street. Another way of looking at it is, Look at how much traffic is left when the students stop using Ricci Street.

How many people have visited Ricci Street and to Learn?

I don't know. Here's how many users sessions there have been. A user session is define as a series of requests from a user-agent (a browser) at an IP number. If more than 30 minutes passes between requests, it is considered a new session. If the user visits Ricci Street, then browses elsewhere, then gets a cup of coffee, and then returns to Ricci Street within thirty minutes of the last request, that is one session.

The 30 minutes is arbitrary. Its advantage is that I have been using it consistently for nine years, so the numbers on the table above show valid trends.

If I take the number of user sessions on the table above, round down to the nearest five-thousand increment, add them up, and multiply by twelve, I get a rough estimate of the number of user sessions over the past nine years: 4,260,000. That four and a quarter million user sessions. To repeat, I do not know how many different human beings that represents.

I think it is safe to say that to the extent that Ricci Street is a publication, it has reached more people worldwide than all the print publications of all the other Medaille faculty put together.

Why do these numbers undercount the traffic?

Visitors print pages or save them to view offline.

An example is library whacking. In other countries where a remote university pays dearly for bandwidth, faculty are encouraged to treat the web like a library reserve. The faculty member gives the librarian the url, RicciStreet.net, and the librarian downloads the whole site ("whacks" it) or a substantial part of it to the university's servers. The faculty member and the students will access it from there, but none of that traffic will show on my logs.

For example, the September 2003 server logs analysis showed that the month's top single user was at a university in Argentina. That one user in four sessions was responsible for almost three percent of Ricci Street's total monthly traffic.

browser cache

Pages are served from the user's browser cache. In September 2003, cached hits as percent of total - 36%.

server space

From 1998 - 2004, I had only 600 MB available. Since then, I have had 3 gigabytes. More server space filled with more files receives more traffic. As that stays steady, traffic growth slows.

incomplete information

Some files on the server are not strictly school-related. Not all school-related traffic, for example, ParksidePlaza.net, is accounted for here. The latter is a little greater than the former.

databases

They need discrete and countable events. On the tables above, Unique users, or visitors, is a count of unique IP (Internet Protocol) numbers. Do not associate users with people. Examples:

A user can be a spider or robot sent from a search engine.

Three students all work at the same company but only one has Internet access. They separately sit at the one computer and request pages from work the same afternoon. One or the other of them is at the computer all afternoon. They are all coming from the same IP number, so they will count here as one user and one long session.

Mary visits once each from work, home, and class all on the same day. That will count here as three unique users and three user sessions.

What percentage of traffic is student traffic?

I estimate 30% student and 70% non-student. However, that's usage, not people. In October 2006, the logs (RicciStreet and toLearn) show 24,064 Unique Users, and we have only several dozen students in the MBA program and several hundred undergraduates using it. In terms of people, the vast majority of visitors to Ricci Street are not Medaille students.

Note in the table above the next-to-last column, the stats for July 2006 when school was not in session. I taught one summer course with a dozen students. The bulk of that summer traffic must be from non-students, and it persists throughout the year.

To the extent that I can separate the student from the non-student traffic, the non-student visitors visit less often and request more pages. Their entering pages are deep in the site and often referred by a search engine.

Parkside Plaza student webs

In October 2000, Medaille's MBA students started maintaining their own webs at Parkside Plaza.

Clearly, there were more unique visitors than there were students. Some of the traffic was from strangers, others from the students' family, friends, and co-workers. For our MBA students, the courses they took with me were their first publications, and they had a lot of readers.

I don't know whether this outcome is worth measuring, but I believe it had a lot to do with the satisfaction students felt with my courses and with the MBA program in general.

By the end of 2001, the students had filled the server space allotted by AIT. I registered ParksidePlaza.net, got an account with a company in Hong Kong, and hosted all the student webs there for two years. Unfortunately, I do not have the server logs, but I can reasonably expect that the traffic above from the Parkside Plaza section of Ricci Street continued at ParksidePlaza.net.

Then AIT began offering more server space for no more money, and I started letting students FTP to that server again in 2004. They soon began filling it, and now I just delete their files whenever I need to free up space for more student webs.

This web, offered in fulfillment
of the requirement in the handbooks of
Medaille College, Buffalo, NY,
Volume IV: Faculty Handbook,
section 4.5.4.3 Faculty Portfolio,
is © 2007 and licensed under a Creative Commons License.

web established: February 2007
page last modified: July 2007
by Douglas Anderson
http://toLearn.net/portfolio/personal/teachinganalytics.htm