Show HN: Tiny weekend project from scraped Wikipedia data: whodiedhere.com | Hacker News Reader

Show HN: Tiny weekend project from scraped Wikipedia data: whodiedhere.com

whodiedhere.com

6 pointsddod13 comments

Awesome work man! Nice to see people making use of DBpedia too, even if it's indirectly. (I'm sure it would be too slow without putting it in your own database.)

Are you going to put the source code up on Github? I'd love to see it, since I'm relatively new to messing around with Node and haven't even touched Socket.IO yet.

I just put it up on Github now: https://github.com/benwasser/whodiedhere

After I built this I played around with some other information in dbpedia to see if I could get anything juicy out of it. No luck so far, but I'm sure there's a lot to uncover.

Thank you for putting it up there. The fact that you have it all tossed in a json file is awesome.

I'm curious, are you familiar with SPARQL, or are you querying DBpedia in another manner? The reason I asked is that I think SPARQL is a cool (flexible, ultra complex queries) but slow way to get data, and I think node's async file reads make it ideal for working off of sources like that.

Instead of learning SPARQL, I wrote my own script to parse from their format into my own. I should probably devote some time to learning SPARQL, but dbpedia's documentation was really confusing for me and I was basically learning it all just looking through the actual db files.

The search result for 'Berlin' includes Isaiah Berlin ---------------------------- Oxford, Oxfordshire, England

Yeah, I wanted to keep things simple and let people search by name, too, even if it's a little outside the scope of the site. I'll run the issue by the board of directors.

Awesome project. Are you releasing the source code?

Thanks! https://github.com/benwasser/whodiedhere

Neat. How are you doing it?

I had to use a couple databases from dbpedia to get all the information (the death location syntax wasn't uniform) to make my own more manageable database.

As for the server, it's just Node.js and Socket.IO

Thanks for the reference to dbpedia. I had been playing around with extracting infobox data from Wikipedia -- damn, I thought that was an original idea, but I guess that was too much to hope for.

Their stuff really isn't very user friendly (or at least wasn't for me) so there's a lot of room for improvement. I'm sure if you made a simple (and useful) Wikipedia API, you would be loved by all.

The people returned in the results would be much more useful if they were clickable links back to Wikipedia.

Edit: also, its a little thing, but please add autocorrect="off" to your input element, the iPad keeps mistakenly auto correcting place names.

Reply on news.ycombinator.com