Tracking Trends
Though this post is about a trend, I did not put it in that category on the blog because I see it as more of a classroom opportunity. The flu season is all over the news this week, and here's a good way to get into it from another angle with students.
How will this year's flu season affect you, your school, your community? The folks at Google announced this week that they have expanded their Google Flu Trends database to 16 additional countries, including much of Europe. (The site is available in 37 languages too.)
They launched Google Flu Trends last year in the United States. They had found a close relationship between how people search for flu-related topics and how many people actually have flu symptoms. They can track the popularity of certain search queries and then estimate the level of flu, in near real-time. Their method can track trends faster than places like the CDC which take days or weeks to collect and release data.
It is not a perfect system. I suspect that a good number of people searching on flu terms are NOT sick but are seeking information - perhaps preventative in nature. Still, their wisdom-of-the-crowds approach seems to track the actual CDC stats pretty closely. Their own analysis of last season shows that Google Flu Trends had a close 0.92 correlation with official U.S. flu data. They do filter out terms that may be popular because people hear about them in the news. For example, "swine flu" is not used since people seem to be more likely to search for that because they want to know more information about it, and not because they actually have H1N1.
This video gives information about how they built their model.
This is a kind of self-mashup for Google since they mash their own search data with da ta collected by other means (at least for corroboration). Last year they had added the data on the flu's progress in Mexico and now they are using aggregated Google search data to estimate current flu activity around the world in near real-time.
Tracking these types of trends and using the data would make a good real-world class activity in the kind of information literacy that is often overlooked. I think that teachers and then their students often get hung up on information literacy and "research" as being something that must result in a research paper. Look at how Google collects data, what it uses, what it rejects, how it validates its results.
Students can understand that there are more flu-related searches during flu season, and there are more allergy-related searches during allergy season, more sunburn-related searches during the summer. They can actually set up their own trend analysis using Google Insights for Search. They can query search and have an accurate model of real-world phenomena.
I did an Insights search for the term "college" over the past 90 days and restricted my search to New Jersey and got these results.
Final note, the Flu Trends data come from Google.org, not Google.com - their .org arm "aspires to use the power of information and technology to address the global challenges of our age."
How will this year's flu season affect you, your school, your community? The folks at Google announced this week that they have expanded their Google Flu Trends database to 16 additional countries, including much of Europe. (The site is available in 37 languages too.)
They launched Google Flu Trends last year in the United States. They had found a close relationship between how people search for flu-related topics and how many people actually have flu symptoms. They can track the popularity of certain search queries and then estimate the level of flu, in near real-time. Their method can track trends faster than places like the CDC which take days or weeks to collect and release data.
It is not a perfect system. I suspect that a good number of people searching on flu terms are NOT sick but are seeking information - perhaps preventative in nature. Still, their wisdom-of-the-crowds approach seems to track the actual CDC stats pretty closely. Their own analysis of last season shows that Google Flu Trends had a close 0.92 correlation with official U.S. flu data. They do filter out terms that may be popular because people hear about them in the news. For example, "swine flu" is not used since people seem to be more likely to search for that because they want to know more information about it, and not because they actually have H1N1.
This video gives information about how they built their model.
This is a kind of self-mashup for Google since they mash their own search data with da ta collected by other means (at least for corroboration). Last year they had added the data on the flu's progress in Mexico and now they are using aggregated Google search data to estimate current flu activity around the world in near real-time.
Tracking these types of trends and using the data would make a good real-world class activity in the kind of information literacy that is often overlooked. I think that teachers and then their students often get hung up on information literacy and "research" as being something that must result in a research paper. Look at how Google collects data, what it uses, what it rejects, how it validates its results.
Students can understand that there are more flu-related searches during flu season, and there are more allergy-related searches during allergy season, more sunburn-related searches during the summer. They can actually set up their own trend analysis using Google Insights for Search. They can query search and have an accurate model of real-world phenomena.
I did an Insights search for the term "college" over the past 90 days and restricted my search to New Jersey and got these results.
Final note, the Flu Trends data come from Google.org, not Google.com - their .org arm "aspires to use the power of information and technology to address the global challenges of our age."
Comments
No comments