Saturday, January 16, 2010

Google Insights: two-letter combinations, part two

Okay, the answers. Last week, I decided to test myself to see which two-letter combinations I predicted would have been Googled most often. Here are the results:

Set #1: 'au' 86, 'aa' 39, 'ao' 19, 'ai' 16, 'ae' 5. I bet au, aa, ai, ae, ao. So not bad all told. I just underestimated 'ao'. I don't know why people search for 'ao' as often as they do.



Set #2: 'be' 78, 'ba' 11, 'bo' 9, 'bi' 7, 'bu' 5. Similarly, it's not that bad except that I underestimated 'ba'. British Airways, I suppose?



Set #3: 'ca' 78, 'co' 50, 'ce' and 'cu' tied at 13, 'ci' 4. I bet co, ca, ce, cu, ci, which just mixes positions one and two. I forgot about my own country, Canada.



Set #4 is as follows: 'de' at 81 well above everything else, 'do' 13, 'da' at 6 and 'di' and 'du' tied at five. My answer of 'de', 'di', 'do', 'da', 'du' errs by, for some reason, way overestimating 'di'. Too much credit to fans of ex-royalty, I guess. I did say it was the most difficult, though.



Set #5: 'eu' 65, 'ea' 22, 'ei' 9, 'ee' 7, 'eo' 5. This time, like before, I was off by drastically overestimating one of them: 'ee'. I'm not sure why, really. Did I think the world was filled with Estonians?

Saturday, January 9, 2010

Google Insights: two-letter combinations, part one

Okay, so this week I'm going to use Google Insights to test my ability to predict. The game is this: I'm going to take five series of five two-letter combinations, as follows. Set #1 will be 'aa', 'ae', 'ai', 'ao' and 'au'. Set #2 will be 'ba', 'be', 'bi', 'bo' and 'bu'. Set #3 will be the letter 'c' + each of the five vowels, sets #4 and 5 will be 'd' and 'e' plus the vowels.

The questions is: which of each will have been googled the most? Not in any particular year or country – just over all. This week I'll post my predictions, and next week the actual ones. Feel free to make your own predictions, though it's easy enough to cheat, of course.

These are my predictions: for set #1, 'au' will have been searched most, then 'aa', then 'ai', 'ae' and 'ao'. For set #2, I say 'be', 'bo', 'bi', 'ba' and 'bu'. For set #3, 'co', 'ca', 'ce', 'cu' and 'ci'. For set #4, 'de', 'di', 'do', 'da' and 'du' (that one is particularly difficult). For set #5, I say 'eu', 'ee', 'ea', 'ei' and 'eo'.

In a week, let's see how well I do. Is there a point to any of this? None whatsoever. But there you have it.

Saturday, January 2, 2010

Google Insights: Mayan calendar



This is the graph generated when you look up who has been looking up the phrase 'mayan calendar' over the years. For me, the important thing is that the search records go back to the year 2004. All of six long years ago...

I'm a bit of a cynical person. The thing is that it bothers me how eager people are to adopt end-of-the-world scenarios. The reasons why I didn't include the year 2012 in this search are (a) it is a year, and not everyone googling it is necessarily googling end-of-the-world scenarios, and (b) it's a movie too. A movie about the silly quasi-prophecy, sure, but moviegoers can't be blamed for having a mere interest in the tired disaster-movie genre (I saw the movie 2012. I didn't understand a word of it). So I chose the phrase 'Mayan calendar', which no one would ever google except as regards the so-called ancient prophecy that didn't seem to exist until a year or two ago. A search confirms it: that's a real j-curve rise from precisely zero. Clearly just a few years back, no one had ever heard of this silly nonsense. And I can predict, looking into the future, that this curve will continue to grow until 2012 itself before fading away when the world fails to conform to Mayan predictions and end.

People intrigued by all this nonsense include: Indians, Filipinos, Americans and Canadians. Nobody in Central or South America. Hm.