Tim Schwartz made clever use of an old indicator dial with his piece Paris, 2007. He is using an embedded system to monitor realtime search patterns, in order to determine which Paris is more popular. The meter is constantly updated to show the latest results. It sounds like a pretty fun project to attempt. Anyone know how to get a realtime feed of people’s search patterns? [via core77]
16 thoughts on “Paris 2007, a popularity meter”
Google trends is still in labs, and they’ve been supposed to release an API for it for about 2 years…
There’s a python module though available @ http://www.juiceanalytics.com/openjuice/programmatic-google-trends-api/
Try http://www.googlefight.com/ for ‘search patterns’.
You can query the trends page (http://www.google.com/trends?q=paris+hilton%2C+paris+france&ctab=0&geo=all&date=all&sort=0) and then get a CSV version of the data on the bottom of that page.
If you sign in with your google id, you can get the trends results in CSV.
He says he uses search results, not people’s search patterns. He probably uses a method similar to “google fight” : http://www.googlefight.com/index.php?word1=Paris+Hilton&word2=Paris+France
His other work is pretty interesting too. He used the archives of the New-York Times to show how people were concerned about war depending on the year, or how the view the US have of the world has changed over time.
with a random number generator
This shows you user the relative amount of user searches for the two Paris’s vie Google Trends
It’s not real-time, but my first idea is that there’s probably a way to use the Google API. I used it recently to do searches on individual terms and tint the text based on the number of hits returned. I think you can limit the search by date and get a somewhat real-time indication of the popularity.
Another option would be to get data off of Google Trends. If you poke around the page you can find a way to get a CSV of the data with a direct link.
Comments are closed.