Part 1 The behavourial Data Science project – The Big Five and Brexit

Can we determine who is pro or con Brexit? Yes we believe we can because you are whom you follow. Read on to learn more on how we are working together with the University of Amsterdam and how we are getting closer to mastering behavourial data.

Some people call themselves introvert in their social platform accounts, others qualify themselves extravert personalities. What happens if you develop datasets of their online behaviour. Whom do they follow? Whom do they connect with? And could this lead to a new way of filling in the so-called Big Five Personalities? With the ultimate goal of being able to select the personality dataset you need for a campaign and predict the outcome or conversion.

These are the key questions in a master research study that we recently started with the University of Amsterdam. A ‘perfect storm’ combination because of the University’s focus on surveys regarding behaviour on a big scale and the enriched database with 350 million active social media accounts of MyScocialDatabase.

The Big Five explored

Let’s introduce the investigator: Joran Cornelisse, a youngster of 24, who is finalising his master on this subject. His playing field: our database of 350 million Twitter accounts. We will follow him along his road of deepdiving into behavourial data, his testing methodology and his conclusions on getting practical marketing insights and – in psychological terms – determining personalities.

So, the Big Five traits of personality are five broad domains which define human personality and account for individual differences:

• Openness

• Conscientiousness

• Extraversion

• Agreeableness

• Neuroticism

Joran started the exploration of online behaviour within the domain Extraversion.

His thesis is ambitious: extraversion can be determined based on who follows who combined with tweets and posts on social media platforms. Therefore, a deepdive into a database with more than 350 million active accounts is a unique opportunity for him.

Trial and error in a Brexit test

People are not posting on social media all the time. Actually, recent stats even show a decreasing trend in posting behaviour. But people do follow others. And they on their turn follow others. Could we classify people based on their online behaviour?

Just to find out how people score we executed this little test on pro and con Brexit fans by composing a train data set of most followed on the UK remaining or leaving the European Union. Joran combined these data with an analysis on extraversion and intraversion traits. The outcome – although not having a representative volume – was promising. Combining this outcome with the Big Five personalities should get us to algorithms and more practicle insights on online behaviour.

So, yes we can is visible and we will continue building up this science by trail and error. This will become an exciting journey of mastering data to the next level. No doubt about that.

Next time Joran will have discovered more insights. Follow us and learn more!