Context This corpus was created during my PhD research at the Institute of Computing, from University of Campinas under supervision of Professors Ariadne Carvalho and Norton Roman. It consists on a crowd-sourcing experiment for annotating emotions on tweets related to the Brazilian stock market. Content We made available the following output from our annotation system: **tweets_annotators.csv**: A list of all annotators and their profiles (without personal information) **tweets.csv**: A list of all tweets available for annotation **tweets_annotations.csv**: All the individual annotations with emotions, also indicating the annotator and the tweet id and the annotation date and time **tweets_stocks.csv**: The final corpus with annotations after considering the majority of annotators, containing only tweets with at least 3 annotations **tweets_stocks-full_agreement.csv**: A tweets_stocks.csv subset containing only tweets annotated by at least 3 people, in which all of them agreed upon the emotions or only one marked as "don't know"