It's written in Ruby and Sinatra, using the idea of collaborative filtering and extract keywords to do simple topic modeling. See the online demo: http://deacon.herokuapp.com/mamibuy/suggest

There are 5 different idea to implement recommendations. You can click on any item and see the change on items recommended to you!

[SERVER] ruby ./app.rb

[Console] irb -r ./app.rb

[RESTART] touch ./tmp/restart.txt

[DUMP] filename = 'pageview.csv' open(filename, 'w') do |f| Pageview.all.each do |pv| f.puts pv.username.to_s + ',' + pv.usercid + ',' + pv.count.to_s + ',' + pv.url end end

[DATABASE backup]

mongodump --db deacon_api

mongo deacon_api --eval "db.dropDatabase()"

mongorestore --db deacon_api ./dump/deacon_api/