Psychometric-demographic predictions from Social Media profiles

Big Data, Data Analytics, Digital Data, Data Mining, Data Science, Machine Learning, and the list goes on! We hear a lot about data today. I think analysing data, finding statistical trends and making predictions from data is useful (and fun!). Of course, we also need to educate ourselves about how data is used and the ethical issues around data use. We need to question who, how & why is the data being collected and how is it being analysed. I actually am fascinated about the IPO (input-process-output) of algorithms, APIs and machine learning that develop predictions about people based on their data.

 

If you use Facebook and Twitter, you might want to check out Apply Magic Sauce. You give the site access to your Facebook Likes & Posts, and Twitter tweets.  The program is developed “…by researchers at the University of Cambridge Psychometrics Centre and builds upon a 30-year legacy of leadership in advanced psychological measurement and computational behavioural science”. The main purpose is to help businesses and consumers personalise their experiences; this should result in a better experience than by the traditional click-view that are commonly tracked on websites (Cookies, yum!).

 

Some of my current results are posted below. I find the ‘Psychological Gender’ differences between Facebook data and Twitter data particularly interestingI believe the difference in my predicted gender highlights social stereotypes (this makes sense if you understand that their dataset comes from a large survey of people that live in our society and the majority obviously ‘live’ and perpetuate these stereotypes). On Facebook I post mainly about my family but I use Twitter as my professional learning network (PLN) so I tweet a lot about gaming, robots, IT, coding (all viewed as belonging to the male domain!).  Is it ethically OK to use these predictions to then directly market things that align to my digital footprint?  How much better might my online experience be if it was built with this data?  Might this further perpetuate social stereotypes? How can we ensure it doesn’t and still use this data in a beneficial way? Fascinating stuff!

 

Based on FB Likes, the program finds it difficult to determine my psychological gender. I liked Minecraft (more masculine) and The Cure (more feminine).

 

Based on FB Posts, my age prediction and psychological gender is closer to my real age and gender. Likely because on FB I post mainly about my kids; I wonder if men who post mainly about their kids on FB get a high female ‘Psychological Gender’…oooh, I hope not but I can’t help but wonder…

 

Based on my tweets I am predicted to be much younger and also male! Must be because of all those tweets about gaming and robots.

 

LinkedIn has an interesting Social Selling Index (SSI) which I discovered today via a LinkedIn connection. Below is some data retrieved from my SSI today. I’d love to know more about how the Industry SSI Rank works.  I’m in the top 2% of the Sales professionals in the Education Management industry and yet I am actually interested in making connections on LinkedIn with people from Education, the IT industry and social enterprises.

Without understanding fully how this works yet, it appears that I am doing a pretty good job at connecting on LinkedIn. I do have questions about the Industry SSI Rank and I wonder if the industry that I’ve been put into could be tweaked/changed.

 

Data like this leaves me with so many questions. Many of the people I connect with are teachers. From my experience with presenting about the benefits of a PLN and Twitter to teachers, I have noted that many are not comfortable (privacy &/or limited digital confidence issues) to network, collaborate and share online. It would be interesting to find out if this is why the numbers are low or is it more about the common view/myth that LinkedIn is only for job seekers.

 

There is likely an algorithm that does similar on Instagram profiles, I haven’t stumbled on this yet so if it exists let me know.  I did find this interesting paper though: http://www.cp.jku.at/research/papers/Ferwerda_Empire_2015.pdf  Imagine what it might mean if we are determining personality traits from reviewing the posted Instagram images and their filter settings. How might this impact employment hiring?

 

Next thing I’ll be trying: Data Selfie  Anyone used this yet? What did you think of it?

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s