Study: Can a Twitter user’s social media “genotype” predict future behaviour?
We are — and will be — what we tweet. A new study by researchers at the University of California and the Rensselaer Polytechnic Institute in New York has drawn inspiration from the world of genetics to establish what the authors describe as a social media “genotype” — a set of behavioral markers that can be employed to predict a user’s future actions.
That blinding flash you just experienced is a million light bulbs going off above the heads of a million marketing directors all at once. But you already knew that would happen — or at least you would have, had you applied the researchers’ methods. Alas, its slightly more nuanced than that.
In terms of its practical application, the biggest promise of the findings relates to predicting how information spreads through the network according to Petko Bogdanov, one of the study’s authors.
He told Which-50, “Modeling the topic-specific user behavior from social media has great promise in terms of predictive power…Does a topic go viral or not? What topic should one accent on in order to keep their audience engaged, or how to engage a future desired audience?”
According to Bogdanov, “When it comes to predicting what a single individual will do on the social media, the challenge is that there is a wide variation in how engaged people are and what roles they play: active contributors versus passive listeners.”
(Image: Petko Bogdanov, University of California)
He said the more a user’s activity is observed the bigger “reads” of their social genes become available leading to an improvement in predictive power.
However, there are limits to what can be achieved. “When it comes to behavior outside of social media there are the additional challenges of to what extent people’s online identity matches their offline behavior. It costs much less to tweet criticism about a government, for example, then to actually go on the streets and protest.”
The research team included Bogdanov as well as Michael Busch, Jeff Moehli, Ambuj K. Singh and Boleslaw K. Szymanski.
The research paper is called “The Social Media Genome: Modeling Individual Topic-Specific Behavior in Social Media” and the research team’s goals included:
- Proposing a genotype model for social media users’ behaviour that enables a rich-network analysis;
- Validating the consistency of the individual genotype model;
- Quantifying the differences of behaviour-based influence backbones from the static network structure in a large real-world network;
- Employing genotypes and backbone structure for adopter/influencer prediction and latency minimisation of information spread.
The authors write, “The social media genotype, similar to a biological genotype, captures unique user traits and variations in different genes (topics). Within the genotype model, a node becomes an individual represented by a set of unique invariant properties.”
While the study might be a little impenetrable to the rest of us, there is an easy-reading analogue available from the MIT Technology Review. According to Technology Review’s summary, the researchers analysed two sets of data: the first consisting of 467 million tweets by 42 million users in 2009, and the second consisting of 14.5 million tweets from 9000 users in 2012. The study analysed content based on hashtags from five topics: sport; business; celebrities; politics; and science/technology.
The conclusion, according to Technology Review, is that “Every person has a fixed set of interests, called their social-media genotype, which determines their pattern of behaviour on networks such as Twitter. What’s more, they say that once these genotypes have been discovered, they can be used to predict an individual’s future behaviour.”
Bogdanov told the web site, “Our hypothesis is that individual users exhibit consistent behaviour of adopting and using hashtags within a known topic.” It was this consistency that allowed the researchers to identify stable patterns of hashtag adoption — basically a social media genotype.
According to Technology Review, “Bogdanov [went] on to identify substructures within Twitter through which hashtags of a certain topic tend to flow. They call these structures ‘topical influence backbones’ and say that they are subtly different from the well-known followee/follower networks that exist on Twitter.” They then discovered the combination of a user’s social-media genotype and their relationship to topical influence backbones enabled predictions about their likely reaction to a hashtag on any a given topic from somebody they follow.