Sentiment-weighted word clouds

[bernardofn]

Expected behavior

Enhance the widget Word Cloud with sentiment-weights.

Actual behavior

Word clouds only represent tokens frequencies.

Steps to reproduce the behavior

It could just introduce a "color by" in word clouds.

Additional info (worksheets, data, screenshots, ...)

Example of output: https://hackernoon.com/using-machine-learning-to-visualize-customer-preferences-6a007cfb9b97 http://www.nytimes.com/interactive/2012/09/06/us/politics/convention-word-counts.html

[ajdapretnar]

@bernardofn Thank you for a good idea! We have discussed this and it turns out it is not completely trivial. We need to think a bit more what would be the best way to implement it. Stay tuned. :)

[bernardofn]

No worries. Just wanted to share with you an interesting visual I recently found.

I will stay tuned :-)

Sent from my iPhone

On 8 Sep 2017, at 09:50, Ajda notifications@github.com<mailto:notifications@github.com> wrote:

@bernardofnhttps://github.com/bernardofn Thank you for a good idea! We have discussed this and it turns out it is not completely trivial. We need to think a bit more what would be the best way to implement it. Stay tuned. :)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/biolab/orange3-text/issues/311#issuecomment-328028942, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AcEu4ceaN2-uSm8psS9u0h243tl5QLTvks5sgPHRgaJpZM4PPzws.

[ajdapretnar]

We have recently discussed this and sentiment-weighted words clouds are not a realistic option in Orange. This is because Sentiment Analysis gives scores per document, not per word. The option that we discussed was to have an option to color by a discrete variable, where the color would correspond to the majority class and opacity to its prevalence.

[ajdapretnar]

Opening a new issue for this one.