Wandora

Hi all,

I'm new to Wandora, and have been trying to figure out how to create a few informations mash-ups of my own.

For example, I'd be interested in importing some Tweets via the Twitter extractor, and then running Alchemy's Sentiment analysis extractor ON the tweets I've just imported into Wandora.

Is this possible? I think it may have something to do with Subject Locators, but I haven't been able to figure it out.

Ideally I want an automatic system where Wandora will input tweets for a given search term, and then run the Alchemy analysis on them.

Any help is much appreciated!!

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

Thanks very much for this! I managed to get it working.

But there's no way to do this automatically - e.g. via coding some kind of script in Wandora?

Great news, you managed to perform a sentiment analysis for Twitter messages.

I suppose one could write a Processing script [1] that would do all the required operations. Processing topic panel is usually used for topic visualizations but the usage is not limited anyhow.

Also, I suppose it is possible to write an R script [2] that does the job.

If you are familiar with Java programming, I encourage you to download Wandora's source code distribution package and program a new tool [3] for the task.

Kind Regards,
Aki / Wandora Team

[1] http://www.wandora.org/wandora/wiki/ind ... opic_panel
[2] http://www.wandora.org/wandora/wiki/ind ... in_Wandora
[3] http://www.wandora.org/wandora/wiki/ind ... r_own_tool