In his explorations with the Twitter button, Tweetmeme founder Nick Halstead discovered that there are millions of tweets a day producing a stunning amount of valuable information, the only problem is that’s pretty hard to separate the wheat from the chafe, the signal from the noise or a bunch of other sifting cliches.
Datasift, in the same space as HootSuite and Tweetronics, is attempting to make this process of sifting through realtime data easier for companies. No longer unique to Twitter, the Datasift platform, accessible through a drag and drop graphical interface, is a curation engine which relies upon realtime filtering, providing developers with alerts, analytics and a realtime API.
Datasift allows mobile, web and desktop client developers to tune tweets through a number for filters, including keyword, popularity, geolocation, and (through a function Halstead calls “augmentation”) sentiment peer indexes like Klout and links. You can start with one rule, and filter it down even further.
For example, Datasift can present you with a stream of the Foursquare checkins at every Starbucks around the world, and then filter them down to everyone with a Klout/PeerIndex score of 50 and saying something positive. This kind of sifting is very valuable to marketers and media companies looking to target very specific campaigns.
While Datasift’s focus on structured data might be too technical for the typical consumer, many businesses revolve around searching through and analyzing realtime information, but don’t have the server resources to filter the Twitter firehose or who don’t want to go through the Twitter app authorization process.
“Datasift is not meant to be an end product,” says Halstead. And his rebranded Tweetmeme targeting developers who need precise streams of data,with a freemium ‘pay for the volume you consume’ or have ads show up in your stream model.
With $1.5 million in funding and the hard lessons learned from the experience of Tweetmeme, it seems like Halstead this time at least has a shot at making all sorts of realtime information a little easier to sift through.
Feedback and Q & A by expert judges Jeff Clavier, Gina Bianchini, Jim Lanzone, Ted Maidenberg, and Chris Sacca. I’ve abbreviated their names, for brevity obviously.
JC: If I want to filter it through my friends, how do I do that with your platform?
Datasift: If you want to set up an alert on a curated list, it’s a platform, so you can use it to anything.
TM: Are you in an arms race with Twitter?
DS: We’re not a client we’re a platform, this is bigger than Twitter, this is about relevant content and push data.
CS: Twitter hasn’t even done search 1.0 yet. It’s not about search it’s about ]push. this world is increasingly driven by analytics. It might not be a huge comany on it’s own.
JL: If you can open this up for people start innovating than I’m all for it.
DS: Tweetmeme processes 1 billion requests. We want to help the people who can make the next Flipboard.
GB: This is a really nice way of bringing it to life.
DS: We already have already 13,000 sign ups from businesses. Tweetmeme is very much a data-driven business.
JC: Can I filter buy # of retweets?
DS: Yes, you can even do it backward and really let your imagination go wild.
JC: Doing this at this scale is hard.
DS: Yes, but it’s viable, 65% of tweets are generated from third party applications.