Finding qualities in requirements-related information is important. Quality requirements are central in building reliable software. However, an isolated identification of qualities is not enough; the impact of an individual quality may compete with another quality requirement. This can be perceived in the example: 'the trade-off between usability and security is not completely secure'. This work studies the use of sentiment analysis to help finding relations among qualities. We will focus on usability and will depart from available NFR (Non-Functional Requirements) catalogues for this specific NFR. The catalogues will be the seed for building a corpus, based on queries in GitHub's Issues. We are aiming to find a list of sentiment expressions that will characterize important relations among usability and other qualities, through the analysis of our mining study. We will contextualize our results in connection with recent work on sentiment analysis for RE, focusing on the specific case of NFRs.