gamebusterz / french-sentiment-analysis-dataset Goto Github PK

How was the data generated?

The Readme mentions:

A collection of over 1.5 Million tweets data translated to French, with their sentiment.

Was the sentiment calculated for English tweets that were then translated to French?

Hi,
Straight forwards, I admit not to have tried the dataset. However, reading your examples as a fluent french speaker, I can tell you the translations are meaningless.

If I translate your phrases back to english:
FR : Awww, c'est un bummer. Tu devrais avoir david carr du troisième jour pour le faire. ;ré
EN: Awww, it is a bummer. You should have david carr of the third day to do it. ;ré

FR: J'ai plongé plusieurs fois pour la balle. A réussi à économiser 50% le reste sort de limites
EN: I have dived multiple times for the ball. To achieve to save 50% the rest goes out of bounds.

FR : Êtes-vous prêt pour votre mojo makeover? Demandez-moi des détails
EN: Are you ready for your mojo makeover? Ask me details

FR: Joyeux 38ème anniversaire à mon livre de tous les temps !!! Tupac amaru shakur
EN: Happy 38th anniversary to my book of all times !!! Tupac amaru shakur
(the last words are not french)

I don't know how this affects the performance of the dataset, but I though you should know