How Big Data Analytics Reveal Your Most Intimate Secrets
Big Data Analytics allow us to automatically analyze digital records such as our updates, photos and ‘Likes’ on Facebook. But did you know that your 'Likes' on Facebook could expose intimate details about you as well as personality traits you might not want to share with anyone? Some things about ourselves we'd rather keep private, right? Most of us don’t openly share with the World sensitive personal attributes such as our sexual preferences, our religious or political views, how intelligent we are, how happy we are with life or whether we consume alcohol, cigarettes or even drugs.
However, a recent study shows that it is possible to accurately predict a range of highly sensitive personal attributes simply by analyzing the ‘Likes’ you have clicked on Facebook. The work conducted by researchers at Cambridge University and Microsoft Research shows how the patterns of Facebook ‘Likes’ can very accurately predict your sexual orientation, satisfaction with life, intelligence, emotional stability, religion, alcohol use, smoking, drug use, relationship status, age, gender, race and political views as well as whether your parents are separated. It is quite scary that those “revealing” ‘Likes’ can have little or nothing to do with the actual attributes they help to predict and often a single ‘Like’ is enough to generate an accurate prediction using logistic or linear regression analyis.
Who would have thought that a ‘Like’ for ‘Curly Fries’ is a strong predictor of a high intelligence - maybe this is a good place to admit that I love curly fries (never thought I would share this publicly). Anyway, here are some of the most predictive ‘Likes’ identified in the study:
- For high intelligence: Curly Fries, Science, Mozart, Thunderstorms or The Daily Show
- For low intelligence: Harley Davidson, Lady Antebellum, Chiq, and I Love Being a Mom
- For Satisfaction with Life: Swimming, Jesus, Pride and Prejudice and Indiana Jones
- For Dissatisfaction with Life: Ipod, Kickass, Lamb of God, Quote Portal and Gorillaz
- For being emotionally unstable (neurotic): So So Happy, Dot Dot Curve, Girl Interrupted, The Adams Family, Kurt Donald Cobain
- For being emotionally stable (calm and relaxed): Business Administration, Skydiving, Soccer, Mountain Biking and Parkour
- For being old: Cup Of Joe For A Joe, Coffee Party Movement, The Closer, Freedomworks, Small Business Saturday, and Fly The American Flag
- For being young: Body By Milk, I Hate My Id Photo, Dude Wait What, J Bigga, and Because I Am A Girl
- For being gay (male): Kathy Griffin, Adam Lambert, Wicked The Musical, Sue Sylvester, Glee and Juicy Culture
- For being straight (male): X Games, Foot Locker, Being Confused After Waking Up From Naps, Sportsnation, WWE, and Wu-Tang Clan
The thing is, when we click ‘Like’ we want to show our friends on Facebook that we feel positive or supportive of specific online content such as status updates, photos or products, books, music or other individuals such as celebrities. What many of us don’t realize is that by doing so you openly share information about yourself that can then be used to predict other, more personal, attributes that you would not share so openly. We now live in a world where everything is digitalized – were we consume music in digital formats, read eBooks, where we shop online and interact with friends and colleagues in social media. This also means that we leave a digital trail of our life and our preferences, which in turn can make it easy to figure out our attributes and personality traits.
Predicting personality traits and attributes is nothing new. For example, personality questionnaires have been around for a long time and they do accurately predict personality types and traits. However, what was different in the past is that we had much more control over the process – we had to complete the survey or give others permission to use our data. With Facebook ‘Likes’ it is slight different because they are by default publicly available. This means that the information you reveal by clicking on a ‘Like’ button can – by default – be used or ‘exploited’ by others using analytics – some with good intentions others with bad ones.
Commercial companies could use this type of Big Data Analytics to dynamically customize the ads you see on your Facebook page (or in fact anywhere) based on your personality traits. Just think of an online ad for the latest car – for people that are classed as shy, reserved and married the ad might highlight safety and family friendliness, while for an single, outgoing and active person it might highlight the attractive design and sporty drive. More worryingly, governments could (and do) use this type of analysis to identify our political views and how they are shifting. Insights from this can then be used to identify how to target election campaigns, etc.
One problem, of course, is that these predictive models are not perfect – no model ever is. Therefore, not everyone who likes curly fries is automatically highly intelligent. The danger is that we might use the insights from predictive modeling to label people wrongly. I can imagine simple mobile phone apps that would allow you to predict personality traits of your friends using this kind of detail. Would you like that? Do you feel that this type of analysis invades your privacy? Will you think twice about ‘Liking’ anything on Facebook from now on? Let me know what you think…share your views…
Bernard Marr is a globally regognized big data and analytics expert. He is a best-selling business author, keynote speaker and consultant in strategy, performance management, analytics, KPIs and big data. He helps companies to better manage, measure, report and analyse performance. His leading-edge work with major companies, organisations and governments across the globe makes him a globally ...
Other Posts by Bernard Marr
The moderated business community for business intelligence, predictive analytics, and data professionals.
|How do you innovate effectively and maintain a competive edge?|
Learn how in our exlcusive ebook, "Bad Data Need Not Apply: Designing the Modern Data Warehouse Environment."