This is a very cool mashup application using time lapsed “tweets”
created by the New York Times to track what was capturing the people’s
attention during the Super Bowl. The tweets are shown on a map of the
US to show where they originated, and like a tag cloud, the size
denotes how many people are tweeting those words.
The NYTimes
folks put together several fun cuts of the data. My favorites are
below, but to really capture the coolness of this mashup, be sure to
“play” each tweet stream.
The reason why I think it’s so
interesting is that by displaying the information in this way, you can
understand the span of the Super Bowl experience, especially in
relationship to events like scoring, halftime shows, and commercial
ads. It’s like watching the Super Bowl in super fast forward, but from
the perspective of thousands of viewers.
This may also lead to
some interesting new analytical tools for businesses interested in
mapping out both geographically and temporally buzz about their
company, product, or service. Mouse over any of the words and you’ll
see a breakdown of the number of people saying each word in each metro
area. It’s actually somewhat surprising how low the numbers are, I
think mostly because few Twitter users turn on geo-tagging of their
posts.
In many ways, I think this simple mashup shows the
power of Twitter — it’s when you see the aggregate whole of Twitter
that you begin to understand the potential for it to serve as a channel
for what people are paying attention to. Take a look at the “talking
about ads” stream to understand which ads resonated where — media
buyers, get your orders ready!
Here’s a screenshot of the “all
‘tweets'” stream during the half time show, where it clearly shows that
we are indeed a country united by Springsteen:
And
this screenshot is the “People saying ‘go'” tweet stream, just after
the last Cardinals touchdown. You have to “Play” the animation to get
capture the excitement of the moment. It was when I saw this animation
that I really understood the power of this application. Really, take
the time to try it out — it will give you a whole different
perspective of the aggregating power of Twitter.
PS: I found out about the map via Twitter, specifically thanks to @biz Stone, CEO Co-founder of Twitter. Kinda demonstrates the power of Twitter, huh.
Update:
I contacted Matthew Bloch and Shan Carter at the New York Times, the
folks who created the Twitter map application. Matthew provided me with
some background on how they did it:
make the Twitter map, Shan harvested the data and I programmed the map
interface. Shan’s Ruby script connected to the Twitter web service and
downloaded all the tweets that matched a long list of keywords. We
thinned out this data set to remove irrelevant tweets and tweets that
we couldn’t match to a geographic location. This yielded about 65,000
messages.
The
Flash map loads lat-long coordinates, timestamps, city names and a list
of keywords for all of these records (it’s a lot of data, but we
managed to compress it to less than 1MB). Then the map projects the geographical
coordinates, aggregates the tweets into regions and tallies the
keywords for each geographical set. Filter by date and theme, find the
top words across the map, scale the words by count and voila!“
