Pretty! This is some of the prettiest data I’ve ever taken a look at. These here are the trends, for a 24 hour period of the frequency of google searches – not how often these terms are searched for, but how quickly thier popularity is rising. I had expected a couple trends, but nothing like this – just look at this beautiful data!
You can already start seeing some trends here – for example, my script failed twice (I’ve since fixed the bug) and for some reason a lot of people who had been searching for pop culture topics (but not other topics) suddenly stopped at 4 AM for no reason, the started again at 5. I’ve already removed a lot of the noise from this version version – items had to occur for at least 4 hours to show up here, but as you can see it’s tough to pick out individual trends. Try this instead:
This is a little better. Topics tend to start high and then level slowly down as newer more interesting news takes it’s place.
Anyway, this is just a rough version – in actuality this is all just a test for the Stream Graph I plan to make with this data. I have been keeping track of whether or not each item is covered by the NYTimes at the time of its search, so that should make the final product even more interesting.
Want to play with the database or this dataset (including the NYTimes info)? Give me a comment below.