Oh man. Google Insights for Search is good fun. I’m supposed to spend this week finishing a number of writing projects. But I spent almost all today running different searches and being basically stunned at how much data’s available through the interface.
I mentioned earlier today that Google makes “related” search information available – there’s much deeper information available through the CSV interface, giving fifty associated terms for most searches. I have high hopes of playing with this data to revive my clustering tools, trying to explore the “freudian web” of associations between search terms.
My friend and research partner Hal Roberts has been playing with the trending functions of Insights – similar to those offered by Blogpulse, but covering a much broader search population. He’s got an interesting comparison of searches for “digital cameras” – cyclical, with spikes around Christmastime, but decreasing in intensity over the years – and “nigeria” – a steady, low-level of interest.
I think what’s most exciting to me (at least right now) is the geographic information available within the system. Andrew Chen has an excellent post, looking at searches for different social media sites and their distribution throughout the US. He observes that well-established sites like Digg are searched for by pretty much the entire US (the northern plains states seem immune from Digg-interest… and, oddly enough, there are lots of folks searching for Reddit in South Dakota.) Other sites, like TechCruch, are primarily searched for by Californians… (The folks at TechCrunch point out that this shows the disparity between searches for “techcrunch” and “techcrunch.com”.)
What’s really fun, IMHO, is using the location features of Google Insight to find out what topics are popular in what countries. Google won’t give you a list of most popular searches worldwide… but they will give you that list for an arbitrary nation, like Australia or Zimbabwe. We’ve had similar information available via Alexa in the past, measuring what sites are popular in what countries. But the search data has the potential, I think, to be extremely revealing.
Hal points out that one of the most popular searches in Nigeria is for “Email Extractor Lite 1.4”, a web-based tool designed to extract email addresses from a large piece of text, probably copied and pasted from the source of a forum page. Actually, variants of this search are the 2nd through 6th most popular searches in Nigeria.
As Hal notes, there’s pretty much no way to explain the result other than concluding that there’s a disproportionate interest in that particular spamming technique in West Africa. Here’s a map of interest in the search “Extractor Lite 1.4”, a spam cluster with a strong concentration in West Africa, but some interest in North America, the Middle East, Western Europe and Southeast Asia.
I’m wondering what other pockets of “undesirable” behavior are mappable via this technique. For instance, searches for “keygen”, a popular site that offers serial numbers and software keys to enable pirated software shows a heavy concentration the former Warsaw Pact nations, with some strength in Southeast Asia as well.
Searches for “torrent” also have an interesting concentration: a heavy concentration in central Europe and Scandinavia. Suspecting a connection to the Pirate Bay, I searched for that term as well, but discovered interest in Pirate Bay is tightly clustered in Scandinavia. Is it possible that interest in torrents in Central Europe, as compared to Africa or the former Soviet Union, is connected to significantly higher bandwith capacities?
There’s no end to the fun you can have with these sorts of maps. Google doesn’t include sexually suggestive terms in the lists that come up when you poll a country, but you can search for those terms specifically and come up with some fascinating maps. A search for “youporn” – an amateur pornography site that now ranks #35 on Alexa’s list of most popular websites, outpacing Flickr, Friendster and Craigslist – gives this map, with an interesting Mediterranean cluster:
Legendary French epicure Jean Anthelme Brillat-Savarin wrote, “Tell me what you eat, and I’ll tell you who you are.” What do we know about a country based on what they’re searching for? Is there something you’d like to tell us, Italy?