Niagarank.es, robot curation for news aggregationJan 30, 2013
Niagarank has been developed upon the experience of the robotic news curators that we launched on 2012 for several customers. The Kisale Cycling news site and the Nestoria Real-Estate UK News aggregator have worked fine, and we decided to expand the experience. Multiply the potential of our backend machine (we call it Robsoc, Robotic Social Curation), applying the idea to create a newssite tailored to Spain, in Spanish: That is Niagarank.es
The site crawls realtime news watching what the people tweets. In 40 distinct channels we have detected and ranked a community of several thousand users, and tracking their tweets in realtime, we also rank the news that sprout out. Once a given link surpasses the threshold of publication, the link is parsed, and we publish a thumbnail, a snippet of text, and we aggregate the relevant tweets in which that news item has been mentioned or commented.
This is an automated process of curation. For instance, these are the numbers of one of those channels (Política in Spain), 7 days of activity, just the past week. We received, 887201 tweets, and analysed 477271 of them (the rest were by people with 0 rank, so to speak). In those tweets, we detected 97759 original tweets that brought a URL within (a link either to Twitter pictura or to something else). In those tweets, there were 41603 distinct URLs, and the machine ranked those in realtime, deciding which were the relevant ones: just 45. The rest, so to speak, was noise. Distilled information, curated from almost a million tweets. And that process is repeated along 40 channels.
We're just beginning with Niagarank. We plan to develop other country versions over the next months, and also content curation projects for customers.
So far, I just presented the initial stages of the tool in the presentation (see embedded below) delivered yesterday at the Master in Digital Journalism at IE, Madrid. The students were interested, it appeared to be, and I'm thankful of their teacher, Enrique Dans, who gave us this opportunity to explain what we're trying to do.