bubble-up is an app for reccomending articles based on user's preferences
built at Complicity - 2013 by
@brianabelson, @annabelchurch,
and @pudo
- Create a postgres database called
news
(or whatever you want) - Create an environmental variable called
DATABASE_URL
with your database path, e.g.export DATABASE_URL=postgresql://brian:mc@localhost:5432/news
- Build the database of newspapers using
build_newspaper_db.py
, which sources data from NewspaperMap - Build the database of articles using
scrape_rss_feeds.py
- Run
web.py
to host the api
I'm hoping to build an intelligent way of determining article similarity via textual features on-the-fly.
my experiments with this can be found in topic_modeling/