Corrected Essays With Scores From UOL

Context

This dataset contains hundreds of corrected essays with scores, all of them which were written according to an monthly selection, and to train Brazilian high-school students to university entrance exams.

Every month, UOL publishes an theme and receives 20 random essays for scoring and commenting.

Content

The content of this dataset is an csv file on which every row contains the essay text and title, the theme text and title, and the scores given.

Inspiration

The motivation for this dataset is to evaluate if we can create an automatic manner of scoring essays. Can we predict the score given an random body of the text in an reliable manner?