Natural language processing has seen a rapid development this decade. It has positioned itself as one of the fields in computational sciences with one with the most promising future ahead of the industry, due mostly to the development of pioneering techniques, but also to the increased affordability of hardware resources that allow a better performance of this technologies. On the other hand, according to the International Federation of the Phonographic Industry, global recorded music revenues grew by 9.7%, and total streaming revenues grew 34% in 2019. Revenue numbers in 2018 go up to US$19.1 billion. It has also been watched there were 255 million users in paid streaming platforms at the end of 2018 [1]. These information shows that music industry is an important industry, with significant revenues. During the development of this End of Degree Project, the goal will be to apply Natural Language Processing techniques to popular song’s lyrics from last years to uncover trends in words use in songs in correlation to the sentiment expressed in the text. Besides it will be developed a recommendation engine to assist the task of composing lyrics for songs. It is also intended to make an analysis of the data obtained by the uncovered trends of words by each sentiment. This project is intended to explore the use of natural language processing tools to assess the viability of these techniques applied to lyrics. The use of this techniques such as bags of words, n-grams extraction or words similarities through word embedding will be investigated. All these tools will be used for both information scraping from lyrics and implementing the recommendation engine. The aforementioned system will be able to show the most common words and n-grams used in songs besides giving recommendation to finish uncompleted verses, everything from a sentiment perspective. This system will offer assistance to the artist to obtain ideas at the time of writing lyrics.