Hits

A little UNIX skill can go a long way:

$ cut -f 1 -d ' ' access_2003-12-07.log | sort -u | wc
    2002    2002   27920
$ cut -f 1 -d ' ' access_2003-12-08.log | sort -u | wc
    1697    1697   23689
$ cut -f 1 -d ' ' access_2003-12-09.log | sort -u | wc
    2146    2146   30011
$ cut -f 1 -d ' ' access_2003-12-10.log | sort -u | wc
    1591    1591   22256
$ cut -f 1 -d ' ' access_2003-12-11.log | sort -u | wc
    1263    1263   17619
$ cut -f 1 -d ' ' access_2003-12-12.log | sort -u | wc
    1318    1318   18364
$ cut -f 1 -d ' ' access_2003-12-13.log | sort -u | wc
     998     998   13991

For the non-techies out there, this means that, for the week starting Sunday, December 7, 2003, I had the following unique IP addresses hit my server:

12/ 7 (Sun) 2,002 Unique IPs (80,584 accesses)
    8 (Mon) 1,697            (60,945         )
    9 (Tue) 2,146            (90,110         )
   10 (Wed) 1,591            (54,584         )
   11 (Thu) 1,263            (40,879         )
   12 (Fri) 1,318            (42,245         )
   13 (Sat)   998            (39,823         )

All timestamps are kept in Japan Standard Time (JST +09:00). Doing a straight word count (wc command) on the logs give the right-most column, the number of accesses. However, the number of accesses include Google-bot and other web crawlers going over the entire site, as well as all of the accesses to the various graphics that adorn each page. This number is really meaningless, except for calculating throughput.

As for the high unique IP hits, that does come as a surprise. (I've never really looked into the matter before.) I see a couple of form-mail attacks, and at least one Microsoft IIS attack to account for a few of those unique hits. Also, I noticed that AOL rotates IP addresses for users while I was investigating a spammer posting a pyrimid scheme to the forums recently. I was quite impressed with such a setup, most likely boosting AOL users' throughput. But it means that for any given AOL user in a particular city, (s)he'll really be using about five to ten unique IP addresses during a session. So, please don't take those numbers as meaning that that's how many people viewed the site on each given day.

While I'm impressed at the high number of IPs hitting the site during the off season, as I've said before, I'd rather have just 10 people view the site who care than 10,000 unique hits a day. (Advertisers, that means I'm not interested in placing your banner ads.)

The focus on players going to the Majors don't excite me, but I still think that about 10 of those hits care about baseball here in Japan. I've been working mostly on data entry lately, so I have a feeling I've been letting those 10 people down by not providing regular news. Sorry.