Crunching the numbers

No thanks to Bluedomino

I thought I would look one last time to see if Bluedomino had come up with the logfiles for March, even though I had heard nothing from them.

surprisedI was shocked to find a huge logfile — three times the normal size! I downloaded the file and started to crunch the numbers.

chokeThe analysis program spewed out error messages, complaining that the data file was "misordered." All the records should have been in order from earliest to latest, but they were all jumbled.

frowningNot to be deterred, I sorted the file by date and time. Since the file contained 152,400 lines of data, it's a good thing I have a tool to do this.

frowningThe analysis program then ran without complaint, but when I looked at the results, it was obvious something was wrong: the number of pageviews appeared to be more than twice those in any previous month. That would be wonderful, but it clearly can't be right.

frowningI sorted the file again, this time deleting any duplicate data — almost 81,000 lines of data. The Bluedomino people must have cobbled together my logfile from backups, without taking care to make sure the records were in proper order and not duplicated.

smilingI ran the program again, and looked at the results. Yes, that's more reasonable.

frowningBut wait! The last four days of the month have no data at all, and the day before that is obviously very, very incomplete.

winkingAha! I can do the math! After calculating the average number of pageviews per day for the 26 complete (I assume) days, then adding on 5 times that for the missing and incomplete days, I came up with as good a total as I'm ever going to get.

It's another new record. When I looked at the referrals (where people were before they landed on my page) I see that I am getting a lot of traffic from Google and Yahoo search engines. The Google users, in particular, seem fond of my Bush pictures.

And although it's not counted in the pageviews, someone from fark.com is stealing my bandwidth, since over 6,600 files were sent without a single pageview being registered. Here's the picture that was hijacked. Since I stole that picture from the NY Times, that makes them an accessory after the fact! My friend Jim gets particularly exercised when people link to his pictures, and he goes to great lengths to substitute an outrageous picture for the one the poster thinks is linked.

See the results