Hi Sentitrack
In this specific case, feeding a sentiment analysis extractor with twitter feeds, the trick is to transform twitter messages (stored as variant names) into occurrences. I assume you have successfully performed a twitter extraction. Next
* Open the topic named as
Tweet. You'll find all extracted twitter messages as instances of this topic.
* Select all instance topics.
* Right mouse click the selection and choose menu option
Topics > Occurrences > Make occurrences with all variant name languages....
* Select topic for occurrence type. Created occurrences will have this topic as a type.
* Wandora performs the transformation. Now all tweet topics have an extra occurrence that contains the twitter message.
* Select all instances of
Tweet topic again.
* Right mouse click to reveal popup menu.
* Select menu option
Topics > Occurrences > Refine > With Alchemy sentiment extractor.
* Pick up both occurrence type and scope topic. Scope is probably
English language topic. Click OK.
* Wandora asks your Alchemy API key. Enter your API key. Click OK.
Now Wandora performs sentiment extractions. Once the operation finishes all tweet topics contain an association to a topic expressing the sentiment type (negative, neutral, positive).
Just to let you know that I just tried to do sentiment extractions to Twitter messages written in Finnish and Alchemy API returned an error. No, it didn't get Finnish. I hope you are trying the sentiment extraction of messages written in English.
And that's about it. Unfortunately Wandora doesn't support automated chains of actions, yet. You have to do all steps manually.
Kind Regards,
Aki / Wandora Team