Traffic Report
other pages
month-by-month trends | visitors' domainsthis page
How many? | From where?
What browser / computer?
How is this information collected and analyzed?
What else? | Learn more
That's as hard to answer as where "out there" is.
The short answer: Medaille College students, of course, because they need the information and contact with other students. What surprised me from the first month (March 1998) and continues to fascinate me is the non-student traffic. The visitors come from over half the countries in the world. The Internet's inherent anonymity makes almost everything else about these visitors a mystery.
Note the month-by-month traffic to the students' Parkside Plaza webs.
It depends what you mean by visits. Take a look at the summary table and graphs of the month-by-month trends. These stats can answer that question only by inference. I assume that:
student
traffic includes all the IP addresses from Medaille as well as from the
students' employers
non-student
traffic includes all the non-U.S. IP addresses as well as those clearly not from
the Western New York area such as universities in other states.
Between those safe assumptions is a large gray area. I have the stats from two summers now when student use dropped way off. Without an elaborate and off-putting password system, however, I don't know how to further separate student use from non-student use.
Then there's caching. You get the page from your ISP's cache and the hits count there, not on the Ricci Street server. Because students are more likely to be recent visitors and to ask for cached pages, they are more severely undercounted by caching. How bad is it? Of all the requests every month, over a quarter of them are labeled 304, a server code which means Not Modified. You get the page from your PC's cache. Back on the server, those requests count as one hit instead of initiating a series of requests that would be however-many-more hits. Learn more about caching.
Based on that large gray area and not counting the administrative and search engine robots and spiders, I estimate 40% student and 60% non-student. However, that's usage, not people because I don't have that many students. 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. Also, I have a pretty good idea of how many students are using these pages, a couple dozen per semester. That leaves hundreds of visitors from elsewhere.
How many hits do other sites get?
Internet.com's Cyberatlas
has traffic counts at the Web's most popular sites.
A little over half of the requests come from .com addresses and another ten percent from .edu, .net, .org, .mil, and .us, usually in that order. Over a third of the requests are from non-U.S. addresses.
Half a dozen countries turn up in the logs every month, usually in this order:
Australia, Canada, France, Germany,
New Zealand, Singapore, United Kingdom
Other countries have missed only once or twice, but I'm throwing them in with the countries that have turned up at least once through June 2000:
Argentina, Austria, Bahamas, Bahrain, Barbados, Bosnia and Herzegovina, Brunei Darussalam, Bulgaria, Belgium, Brazil, Cayman Islands, Chile, China, Colombia, Costa Rica, Croatia, Cyprus, Czech Republic, Denmark, Dominican Republic, Ecuador, Estonia, Finland, Greece, Guadaloupe, Guam, Hong Kong, Hungary, Iceland, India, Indonesia, Ireland, Israel, Italy, Jamaica, Japan, Jordan, Kenya, North Korea, South Korea, Kyrgyzstan, Latvia, Lebanon, Liechtenstein, Lithuania, Luxembourg, Macau, Macedonia, Malaysia, Malta, Mauritius, Mexico, Morocco, Namibia, Netherlands, Nicaragua, Norway, Oman, Pakistan, Peru, Philippines, Poland, Portugal, Romania, Russian Federation, Saudia Arabia, Slovak Republic, Slovenia, South Africa, Spain, Sri Lanka, Sweden, Switzerland, Taiwan, Thailand, Trinidad and Tobago, Turkey, Ukraine, United Arab Emirates, Uruguay, Venezuela, VietNam, Virgin Islands, Yugoslavia
It's a lot like reading license plates and bumper stickers while on a cross-country drive. Or puzzling over the stickers on grandpa's old steamer trunk. Or thumbing the passport of an inveterate traveler. If such appeal to you, you're welcome to pore through July 1999's list of all the top couple hundred domain names.
Generally, the .com, .net, and .org addresses are registered to post office addresses in the United States. Other countries have addresses that end in 2-digit country codes such as sg for Singapore or au for Australia.
Most recent: Microsoft Internet Explorer 60%, Netscape Navigator 30%, others 10%
Trends: This 60/30 split reversed from 1998 to 1999. Into June 2000, it was holding steady. I assume that's because Netscape hasn't released a fifth-generation browser and Microsoft's is more standards-compliant.
Since the world-wide trends are showing 85% Microsoft, the lower numbers here must come from the greater number of students using Netscape in class.
As the third-generation browsers become rare, expect the fifth-generation browsers to become common. Because few pages look good on all browsers, these statistics tell me I can concentrate on the latest versions of the mainstream browsers.
Other browsers: Dozens of different browsers are in the logs every month. The most popular, in roughly this order, representing way under 1% of the hits each: Architext Spider, Googlebot/1.0, libwww-perl/5.33, ia_archiver, Rex Swain's HTTP Viewer, W3C_Validator/1.62 libwww-perl/5.43, Teleport Pro/1.29, Lotus-Notes/4.5, Mata Hari/2.00, www.logika.net, Lynx.
Most recent: Almost all Windows 98, Windows NT/2000, and some Windows 95. Macs fluctuate between 5% and 10%. Unix 1%.
Trends: No surprises here. Almost all the Mac visits come from Medaille's New Media Studio, one of Ricci Street's proprietors, and a couple of students. The hand-helds and WebTV started to appear in spring 1999. I expect their share to rise. I also expect Windows 2000 (NT) to continue to rise.
The Apache server tracks all the HTTP requests in log files that I download once a month. Until December 1999, I ran the database through OpenWebScope and cut-and-pasted the results onto these Site Stats pages. From January 2000, I've been using Web Trends. I would be more than happy to share the raw database files, especially if you can tell me something about them that I don't know.
The stats up to summer 1999 are for the toLearn.net domain. After that, RicciStreet.net and ClearLightStudio.com became available. As of September 1999, toLearn.net has:
my
personal stuff, including tests of new pages I'm developing
the old
(1998-99) MBA 604 web
John Donovan's
math course webs
The aggregate stats reported here exclude the personal stuff and the test pages, but include MBA 604 and John's math pages as well as all the traffic on RicciStreet.net and the affiliate sites, ClearLightStudio.com, Little-Lavatelli.com, and Andersonsbooks.com. I estimate that less than 3% of this reported traffic is administrative -- the person who made the page checking looks and links. In addition, regular visits from search robots account for less than 5%. On the other hand, because of caching, the server logs undercount, but I don't know by what percentage.
However, the administrative and search engine traffic as well as the caching are constant and stable. Thus, these stats are comparable month to month and the trends are valid. If you think otherwise, I'd appreciate hearing from you.
terms used on the Traffic Report data tables
browser | cache | hit
| request | server | server
log
server log entry | visit, visitor
How much should webweavers know about you?
|
|||||||||||||||||||||||||||||||||||||||||||||
