Using Analytics to understand the summer hit Despacito.

What is the hit of the summer 2017? I’m pretty sure that from the “botecos” of Rio de Janeiro to the coolest discos of Barcelona and the indoor malls of Milano Despacito is the number one!

The remix of Luis Fonsi’s Latin crossover hit, featuring Daddy Yankee and Justin Bieber, had scored for weeks the Billboard Hot 100. It’s also helped Daddy Yankee become the first Latin artist to be named Spotify’s top artist.

I’ve been analyzing Latin summer hits in the past and for most of them, the secret for the song’s success was obvious for me: the slinky rhythm and the simple and infectious hooks delivered by a really easy to learn chorus added to a catchy and upbeat sound (much like a lot of Spanish music). Put everything together, shake it and so it becomes popular, for sure.

In my opinion, another key factor in the success of those songs is actually the collaboration with Latin urban acts and the fusion of pop star like Bieber and urban itself, for example: do you remember the hit by Lumedee and Busta Rhymes, I never gonna leave you?

Personally, I’m not a fan of #despacito but I have to bring it down: my daughter love this song, my girlfriend love this song.. and recently I’ve been surrounded by this summer ear-worm.. so I could not ignore it.

Despacito and IBM Watson

This week I started to consider this song as topic for one of mine cognitive analysis using some IBM Watson services and tools to make practice and enhance my skills with cognitive.

So I decided to use Watson Analytics for Social Media to gather some unstructured data on social networks and use Watson Tone Analyzer to perform a linguistic analysis to detect and interpret emotions, social tendencies, and language style cues in the lyrics of Despacito.

With this analysis I was able to measure the incidence of joy, fear, sadness, anger, disgust, confidence, openness, conscientiousness, extroversion, agreeableness and other nuances among the English translation of the song.

The Watson Tone Analyzer Service helped me analyze the lyrics at the document level and at the sentence level.

In this study I’ve been using document level analysis to get a sense of the overall tone of the song and using sentence level analysis to identify specific areas of the song where tones were the strongest.

Analyzing the sentences of Despacito

IBM Watson Tone Analyzer help me to identify sentences with stronger tones in context or sorted by score. Highlighted sentences indicated the likelihood of a tone present.

When more than one tone was detected, the stronger one was shown. You can see here some examples of the Sentence Analysis:

Despacito and the Social Networks

With Watson Analytics and its smart data analysis and visualization I was able to produce a quick social network insight about #despacito using a natural language dialogue and interact with data conversationally to get answers that I could understand.

After create a data-set of 24.650 documents from twitter, web reviews, YouTube, blogs and other social network sources I’ve started to build some visual information charts regarding the #despacito in the range from January 2017 to July 2017.

For example, as you can see in the following picture, I was able to analyze the global “sentiment”about Despacito on social networks, with a 20% of Positive Sentiments, 10% of Negative and a huge majority of Neutral (I am probably included on this huge majority).

Regarding demographics, I learned that on a set of 24k documents, 4716 (19%) were produced by women and 5183 by men! Almost parity.. i would say!

When i comes to understand how the social engagement was divided by countries and cities, I can say that I’ve been really surprised about how Poland (I live here in Poland) has been active during the last month regarding despacito. I think my girlfriend has been tweeting a lot using the hashtag #despacito during the last months :-)

For sure, there would be thousands of other analytics aspects that I should consider to analyze this musical phenomenon using the great cognitive power of Watson but this one was a simple and quick experiment. I will keep an eye on Despacito until the end of this summer and let’s see what will comes out.

For sure the circumstances and mechanisms that produces a summer hit like this are less exciting than its dancing results: a more varied pop landscape where a song can explode out of one market and into the global mainstream.

Songs like Despacito will be lighting up our summer poolside playlists this year. Right now Latin pop is the main benefactor, but anything is possible in the music market: 2018 could be the summer of Ukrainian rave trance or Polish Electronic wave. Does anybody there remember Michel Telò and his “Ai se eu te pego”? God help us!

Probably I will create a Watson Analytics setup to try to predict the trends for the next summer hit and to learn more about the power of cognitive and IBM Watson!!! But this will be another post!!!

See you next time!

Jair Ribeiro

Wave Project Manager at IBM Poland