4

This was a one day project :
I created an app that would directly read feeds about our travel website when they hashtag about their experiences on twitter ,and automated it to pass it through a very minimal machine level algorithm to identify the sentiment of the tweet. (Good ,bad and neutral). The analyzer was about 40 percent accurate,but it did better with training the keywords.This not only helped the Global Communications perform better at their work,but were able to close out most of the issues on a day to day basis.

Comments
  • 0
    Very interesting, how long did it take to build this? :)
  • 1
    @heyheni ,actually,the machine level algorithm was already there and it needed some training on the keywords. Only Wrote an endpoint to that so my feeds can send a text and receive an integer.
    We spent about a weekend for these.

    Getting tweets from twitter took me about half a day and the restful call to call the endpoint about 1 hour. A very basic grails UI for the demo about 1 hour.

    So all in all 3 days.( weekend and the day of the hackathon)
  • 1
    Nice! :D
    So you could do it better now?
  • 1
    @heyheni ,yeah. But there is another team that is maintaining it as it became a pretty good project in the organization. I have moved on from that to something else,but I do follow the updates.
Add Comment