Realtime search has come a long way from just a year ago when the only option really was Twitter’s own search engine. Now Google, Bing, and a gaggle of realtime search startups all have products up and running (even Facebook is expanding its own realtime search to include everybody’s public stream). Today, one of those realtime search startups, CrowdEye, released a bunch of new features that point the way to a better realtime search experience. In particular, it lets you filter your search by location and sentiment (expect Google and Bing to copy some of these soon).
CrowdEye is part of an early group of startups who recently got access to Twitter’s full firehose of Tweets. Today, it turned on that firehose and is indexing a full 14 days of everything on Twitter. CrowdEye returns both links and Tweets as results, and lets you sort by relevance or time. It uses its own CrowdRank algorithm to come up with the most relevant Tweets.
But the best new features are the ones that let you sort by location and sentiment. The location filter shows only Tweets that come from a particular city or spot on a map (you can literally just place a marker on a drop-down Google Map to choose the location). The sentiment filter is more rudimentary. It only allows you to boost positive or negative results, but it doesn’t show you the overall sentiment for a keyword.
CEO Ken Moss says he hasn’t built that yet, but it shouldn’t be too hard and could show how sentiment changes over time. Brands would love that, as well as the ability to get an overall sentiment score by location. With CrowdEye’s new features, you can already filter by location and sentiment and read through the stream of results. For instance, here are negative Tweets about Pizza Hut in Chicago (“Why does pizza hut smell like this ? How can people really ignore this odor ? Why ? Why ? Why ?”), and here are positive Tweets (“Just used the Pizza Hut app for the first time. Overall, great UX and pretty good pizza to back it up. Will definitely use it again.”). Oh yeah, brands are going to be all over realtime location search.
My one compliant: There are no links back to each individual Tweet, only back to the Twitter account it came from.