Deloitte’s closing part at Colours of Ostrava: Textual analytics revealed what performers sing about. And we know what generation Z wants
Colours of Ostrava, day three, the weather is much better. And a workshop that showed a much wider use of textual analytics than you would expect in real life. The last session held as part of the Meltingpot was dedicated to Senta Čermáková and generation Z. The final report from Deloitte’s premiere at the music festival is here.
Did you know that data analytics can also be used in other, more relaxing matters than in business? Such as analysis of song lyrics? Michal Šusták from the Deloitte Analytics team and Matěj Míček, a trainee in our Innovations Team, set upon a journey through the history of song lyrics by analysing five hundred of the greatest Czech hits from the sixties to the present day. The workshop Music in Data: What do performers at this year’s Colours of Ostrava sing about revealed what words appear most often in the hits from the last century and what bands you should see performing if you have children with you.
GET A BETTER UNDERSTANDING OF COLOURS:
Video – Meltingpot and Colours: Recapitulation of the first day
Video – Meltingpot and Colours: Recapitulation of the second day
Video – Meltingpot and Colours: Recapitulation of the third day
Video – Meltingpot and Colours: Recapitulation of the last day
More reports from Colours of Ostrava (in Czech) available.
However, let’s start from the beginning: You have certainly heard about data analytics, so you know that it became very popular thanks to the development of social networks and that it may be used to improve sales, reduce customer churn and also for predicting which employee intends to leave the company soon. You don’t need a crystal ball, but you can’t do without data analytics.
“Companies use these services when they want to know what their customers talk about, what problems they have, how they compare them with competitors and what references they express to the individual products. Analytics can be used in any area where there is a large amount of written text,” Michal Šusták, consultant at Deloitte Analytics, explained. And then he went on to inform the attendees about the conclusions drawn based on data analytics.
Understanding texts in deep: What needs to be prepared before data analysis starts
At the very beginning, there is data – text. In order to be able to work with data, a process called tokenisation must be performed, in which the text is divided into the individual words and meaningful sets (sentences, stanzas, verses – when talking about song lyrics). Then, the base forms of the words must be extracted (infinitive forms, nouns in the first case) and any words that do not bear meaning (such as conjunctions, prepositions, numbers) must be filtered out. Subsequently, words that do not make sense when standing alone are linked to words with which they form a meaningful phrase (good – day). Thus, a sequence of words for a stronger meaning is created. Then machine learning is applied in order to gain clusters, classify the documents, extract entities and topics and finally, the results.
Textual analytics processes unstructured text data (a written piece of text, pictures, sound) gained from communication on social networks, webs, e-mails or call centres. It is often used for creating predictive models in combination with structured data.
From the sixties to the Colours or Songs under the Scrutiny of Textual Analytics
What do you need if you want to immerse yourself in music history, find song lyrics almost 60 years old and analyse all the data? You must spend hours and hours in the archives, in printed magazines, look up contemporary song charts and select the chart-topping Czech hits. Then you must scan them, convert them to text, and let data analytics do its job.
However, Michal and Matěj went even further. From the five hundred analysed songs, they were able to identify whether the meaning of the song was negative or positive or which interpreters belong among the so called kid friendly – their lyrics do not include vulgar language or any other “improper” words, so they are appropriate even for children’s ears. Which Czech band at Colours is the winner? MIG21!
The sixties: the top of the charts is occupied by the song Včera neděle byla sung by Pavlína Filipovská, and Waldemar Matuška’s song Písnička pro Zuzanu. And here is the list of the five most frequent words: you – have – I – peace – angel.
The seventies: Karel Gott, Golden Kids, Holubí dům and songs with the most frequent words: I – you- kiss- love- star.
The eighties: here we go: Holky z naší školky, Nonstop and Zvonky štěstí and the most frequent words … I– you – love – star – satellite.
The nineties: the band Lucie sings their song Amerika, Lucie Bílá introduces her Láska je láska. Surprisingly, love does not attack the first ten most frequent words. The five most frequent words are I – have – go – you – only.
New millennium: Petr Hapka has his song Kocour se schoulil na tvůj klín, the band Chinaski have their song Tabáček, the band Kryštof their Obchodník s deštěm. In addition to the usual words I and you, further frequent words include have – want – know.
Love, love …. and love again, however, sometimes you must also work like a horse (“makat” in Czech) or have gold
What are the most frequent words revealed by data analytics? The Deloitte Analytics team found out that in terms of frequency, the word “you” came before “I” in the sixties, which did not occur again at any later period. And whereas after the normalisation period words such as darkness – live – die were very common, the songs of the following decade reintroduced peace, angel and sky. Then there was a period full of positive words such as lips- love- star and in the eighties, words borrowed from English and newly coined words – such as satellite, song, and “makat” (a newly coined word for “work hard”) appeared often in songs. The change of the regime brought to light new words such as gold, strength and power. The new millennium did not bring the end of the world but a woman (in the top ten), and after a long time, words like God – angel – heart reappeared in songs.
What words are most frequent then? You might be able to guess without analysis – love, of course. However, it is not bad to have your guess supported by data. In this case, the analysis covered more than 25 thousand songs with almost 5.5 million words.
Generation Z in leadership
On Saturday, the last part of this year’s Meltingpot at Colours of Ostrava was crowned by Senta Čermáková who held her lecture on generation Z, which she leads in her Innovations Team at Deloitte. One of the most frequent topics today is robotics, and the advantages robotics can bring not only to work life. However, is there anything that cannot be easily automated? Senta thinks it is people management: “Managing people is a kind of art, you must select the right people at first, then you must get to know them well and motivate them. In the future, professions requiring emotional intelligence will be sought after very much. This is the area where robots will be lagging behind for a very long time.
And a piece of advice in conclusion? In Senta’s guide to better leadership you will get seven straight away:
- First of all, select the people for your team very carefully
- Let them solve problems on their own
- Motivate your people to use their unique abilities and strengths
- Question status quo and motivate your team to do the same
- Create inspiring experience, give them diverse experience
- Develop a wider perspective
- Express and support daring ideas