Lonely Planet Labs

BlogSherpa – end of sprint 3

Johnny CussenLonely Planet author

Hi, I’m Johnny Cussen, and I’m in charge of the technical development of BlogSherpa.

We’ve just come out of sprint 3 on BlogSherpa and we’re now tackling front end integration. BlogSherpa is a blog post syndication service. Bloggers who sign up to the propgram can explicity syndicate individual posts to Lonely Planet. We’ll publish that post on lonelyplanet.com and the blogger gets their own Google Adsense ads on the page where the post is displayed. If someone clicks on an Adsense ad on that page the blogger gets 100% of the revenue. Simple.

We’ll also link to posts from our “Destinations” pages. So on the Rome page there will be a number of links to the most recent posts in the system about Rome. We’re keeping contextual targetting very simple for our beta launch. If a blogger shares a post with us and provides the destination and the country names as tags to their post we’ll match it to the destination.

We’ve also leveraged Reuters’ OpenCalais auto-tagging API . I really like what they’ve done with this simple service. Here’s a description in their words:

“The Calais Web Service automatically creates rich semantic metadata for the content you submit – in well under a second. Using natural language processing, machine learning and other methods, Calais analyzes your document and finds the entities within it. But, Calais goes well beyond classic entity identification and returns the facts and events hidden within your text as well.”

We’ve focused in on the geographic entity extraction capabilities at first and have had promising results. Any tags provided by OpenCalais are identified as such in our content store and we’re working on setting the right relevancy score thresholds. Blogger provided tags are given presidence over any machine tagging. We are seeing some false positives and some geographic features we know about aren’t being detected but overall the OpenCalais service looks good. We’ll be looking for feedback as we trial the service in our beta launch blogger partners.