Until now, it has been seen how extracting information from Beatles WikiPage and Lyrics from Beatles Official page has allowed us to build a network connecting songs depending if their corresponding lyrics were related. Moreover, some communities have been found with regards to that and some interesting conclusions have been extracted. A lot of effort has been put also on performing Sentimental Analysis. This has allowed us to arrive at a bunch of promising conclusions, and offers a new perspective of analyzing a music group trajectory by performing Sentimental Analysis of their songs.
However, there are many more sources of information with respect to Beatles, that as further work could be extracted and combined with the already analyzed ones. One of these sources is Spotify, which offers a lot of extra information of every track such as danceability, energy, key or mode among others. Another source that could be stimulating to use is Open AI. OpenAI is an AI research and deployment company whose mission is to ensure that artificial general intelligence benefits all of humanity. You can read more about their mission and goals here. The idea behind that is to take advantage of this powerful to generate new lyrics, based on already existing ones. Below, these two possible lines of further work are more extensively explained.
Some ideas that can be extracted from this data is to observe how of the added attributes relate one with each other in Beatles songs. Is an increase in energy accompanied with an increase of danceability in Beatles songs? Does an increase of valence imply an increase of danceability?
To solve that some density contours as the ones shown below can be made.
As lines of further work, it seems interesting to relate Sentimental Analysis with Spotify features. Does more danceability implies more positive lyrics? Moreover, a PCA of all the different attributes could be performed in order to observe if some clustering between communities exist in terms of Spotify attributes.
With open AI we can generate artificial text, depending on a certain prompt and other characteristics as input. As a mode of example we have introduced the lyrics of the I Need You song as a prompt, expecting that the AI continues the lyrics.
This opens a new world of possibilities. Could we use this tool to continue the lyrics of a specific song we like? Could we write a "mean song" given the lyrics of a set of songs? For example, taking the lyrics of all the songs in a community and trying to observe how the average song lyrics would look like.