‘Big data’ is the idea that the Internet provides organisations with an unprecedented amount of data that deserves new forms of analysis. The more general idea that sophisticated analysis of big data sets is important is quite topical, just ask Nate Silver.
Google have always been good at this sort of thing. They make (some) of their internal ‘big data’ publicly available for all of us to have a geeky play with. One example is Google Trends which shows (normalised) data for the volume of searching on particular search terms. They also split data geographically.
I’ve been playing around and there are some interesting patterns that suggest to me there is a seasonal flow of geological learning. Take a look at this.
[trend w=”600″ h=”400″ q=”Subduction” geo=””]
There is a clear seasonal pattern there. Searches are lowest in the Summer (July/August) and show a clear dip in December. The pattern holds for plate tectonics and other terms too.
[trend w=”600″ h=”400″ q=”continental+crust, oceanic+crust” geo=””]
It seems pretty obvious these correspond to term time in schools and colleges and that most searching for geological terms is done by students. Further evidence comes from Google Correlate, which shows searches terms that are correlated in time. Subduction searches in France and the US correlate with other terms that students with geological assignments are likely to type in.
Some geological words have other meanings of course, look at this.
[trend w=”600″ h=”400″ q=”quartzite” geo=””]
The peak in January confused me, until I looked at the related terms and found out about the large mineral fair in the town of Quartzite, Arizona in January every year. There is a peak for searches for ‘Mars’ in March which is related to that being the French word for March.
My examples are fairly trivial perhaps, but what staggers me is the potential of this sort of data. Another post I’m writing is about recent scientific research based on publicly available ‘big data’ (remote sensing and altitude data). There must be proper scholarship to be done using this Google Trends data, but for me its just an interesting diversion. I leave you with a puzzle, why is the seasonality of ‘photosynthesis’ searches so distinctive? What happens in October and November? Do thousands of biology teachers all start the academic year with the basics of plant biology?
[trend w=”600″ h=”400″ q=”photosynthesis” geo=””]