Dataset from the paper
https://web.eecs.umich.edu/~mihalcea/papers/ruan.acl16.pdf with original annotiations + added 0/1 mapping (<=0 -> 0, >0 -> 1) and 3 class mapping: [-3,-1) -> label 2 (pessimism), [-1,1] -> label 0 (neutral), (1,3] -> label 1 (optimism)
All credit goes to authors
Finding Optimists and Pessimists on Twitter
Xianzhi Ruan, Steven Wilson, Rada Mihalcea
54th Annual Meeting of the Association for Computational Linguistics (ACL), 2016
