“What,” you might have occasionally wondered “does Mark Gibbs have in common with Penn Jillette?” That’s a fine question, dear reader, and the answer, at least according to Your Celebrity Match, a cognitive computing application based on IBM’s Watson Developer Cloud, is apparently “not a lot”.
Surprisingly, the celebrity whose personality most closely matches mine according to the app is, rather surprisingly, that of Slash, the lead guitarist of the rock group Guns N’ Roses and whose Web site is currently 404'ing on the home page (Time magazine’s 2009 Ten Best Electric Guitar Players of All Time ranked him number two, behind Jimi Hendrix and above Clapton, on a list without Django Reinhard!).
So, guitarist rankings aside, how does IBM's Your Celebrity Match quantify personalities? Driving the app is the IBM Watson Personality Insights service which performs a linguistic analysis of the language used in a public Twitter feed which the Watson-derived system interprets in the context of the Big Five personality traits. It then compares the results to those generated from the Twitter feeds of over 100 celebrities.
The IBM Watson Personality Insights service “[infers] cognitive and social characteristics, including Big Five, Values, and Needs” and what a job it does; on the Watson Personality Insights home page I entered 5,209 words from some of my recent Gearhead posts and this is what I got. I leave it as an exercise for the reader as to whether Watson is right about me …
To use IBM Watson Personality Insights you simply make a REST call to the service’s API providing a minimum of 1,000 words written by the person you’re analyzing (2,000 is recommended) in JSON, Text, or HTML format and the service will respond with a “tree of cognitive and social characteristics in JSON or CSV format.” Pricing for the Standard Service is free for the first hundred API calls per month with additional calls at $0.60 per call and "includes English language and three personality models: Big 5, Values, and Needs."
You can apply IBM Watson Personality Insights to any type of text from tweets, through Facebook postings, to email, blog postings, and documents. I’ve tested a few similar services over the last year or so but Watson’s is by far the most exhaustive. That said, there’s so much data available from an IBM Watson Personality Insights analysis that’s open to interpretation that you’re going to have to be very selective in what you use and careful about how you use it.
I’m still surprised, and slightly puzzled, to find that I'm more like Slash than Penn Jillette. Then again we, Slash and I, are in good company with Neil deGrasse Tyson, Jeremy Gutsche, Nancy Pelosi, Mike Bloomberg, and Rob Dyrdek. Sorry, Penn.