Hi Michael, Tom, Interesting product/project I added a few comments inline below. Cheers.
>> I have a question about using DBpedia properties. >> >> I'm working on an iPad app for visually browsing Wikipedia with radial >> tree layouts. My colleagues and I published a proof-of-concept app >> called 'WikiNodes': >> http://itunes.apple.com/us/app/wikinodes/id433834594?mt=8 >> >> We want to improve the layout and content of the app using DBPedia data. >> For example, for a primary node "Paris", a child node could be >> "Birthplace of", and then we'd list ~10 of the people with >> dbpedia-owl:birthPlace. > > Is there more than one property used for this information? That > sounds like a very low number of instances. By comparison, Freebase > has 3583 people born there (which is still probably just a tiny > fraction of those in Wikipedia). > By experience with similar projects/products at Yahoo! , you don't want to show all the possible related entities, but only the most important ones. Ranking them is important so the user is not overwhelmed. Also, you have limited space or displaying them. That's probably want was meant here with this selection of ~10 persons born in Paris. >> Where/how can we find some guidance on DBpedia properties, and whether >> to use them? -- e.g., a list of all the properties, their frequency of >> use, what they mean (if it's not obvious), and whether they are >> reliable or buggy? Does this exist, or should we extract statistics from >> the downloads? > > I can't help with documentation, but one way to gauge reliability of > properties would to see how well DBpedia and Freebase correspond for > Wikipedia-based entries, since they're basically starting with the > same source data. Where there's large discrepancies, you'd probably > need to dig in to it further to figure out which one is more > reliable/representative. Inventorying and computed these statistics directly from the DBpedia dumps is fairly straightforward, and works pretty well. As for reliability, you can scour the Web for confirming evidence (ie use a web search engine and look for matching phrases), use editors (eg use Mechanical Turks to review facts, or sample of them), or compare with other factual knowledge bases (as suggested by Tom). > >> Also, if this project catches the interest of any list members, I can >> tell you more about our project, as we'd welcome any insights, ideas or >> input on how to use DBpedia data for an app for the general public. > > I'd be a lot more interested in your had an Android, Linux, or Windows > app that I could play with on one of my devices. :-) Count me in! > > Tom > > p.s. Is Wikipedia really 17+? How does Apple enforce this on that > tiny portion of the web that they don't control? ^^ Good question here from Tom! Using a adult/profanity filter would help :) -- Nicolas Torzec. Yahoo! Labs. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Dbpedia-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
