O Twitter, por natureza, é uma mídia social que nasceu com caráter de ‘termômetro’, ‘radar’ dos usuários.
Hoje, evoluiu para um microblog onde o usuários exercem o poder de síntese em 140 caracteres, buscando chamar a atenção e cativar seus seguidores (followers) através de drops de pensamentos, ações, demontrações de inteligência (ou não), informações, notícias etc etc etc.
É, sem sombra de dúvida, uma excelente ferramenta para acompanharmos e medirmos o ‘mood’ dos nossos usuários, consumidores, clientes.
De olho no boom de redes sociais que vem acontecendo há algum tempo, começaram a aparecer diversas inciativas para mensuração e monitoramento desse universo.
Apesar de requerer ajustes, um exemplo bastante interessante está no artigo abaixo publicado pelo Daily News, onde dois cientistas desenvolveram uma forma de analisar o grau de ‘felicidade’ dos usuários através de seus posts em blogs e no Twitter.
Scientists analyze Twitter, blogs, to learn how happy people are
BY Alexandra Hazlett
DAILY NEWS WRITER
August 4th 2009, 11:46 AM
Two scientists from Vermont have created a device that measures happiness by analyzing blog and Twitter posts, MSNBC reported.
The “hedonometer,” created by Peter Dodds and Chris Danforth, combs blogs and Tweets looking for sentences that begin with “I feel” or “I am feeling.”
The next words in the sentences were rated on a happiness scale of 1 to 9. In total, 1,034 words were ranked. “Triumphant” registered an 8.87 on the high end of the scale; “hostage” came in with a score of 2.20 near the bottom.
The scientists used 10 million of these sentences over the past several years to calculate the daily level of happiness.
Vacation days and holidays are among the most consistently happy. President Obama’s inauguration and the most recent presidential election also were high on the list of happy days. On those two occasions, people tended to use the words “pride” and “proud.”
Sept . 11 and often the days before, as well as the recent death of pop icon Michael Jackson were among the saddest days.
Analyzing Tweets, the 140-character messages sent out by users of Twitter, provides for more local insight than blog analysis.
“There are something like 1,000 tweets a minute,” said Dobbs, a researcher at the University of Vermont. “That’s really a lot of data that we can analyze.”
Before including Twitter in the analysis, the software collected information from 2.3 million blogs.
The upside to this method of analysis is that it removes the on-the-spot bias associated with traditional man-on-the-street surveys. When a person randomly polls passersby and asks, “Are you happy today? Yes, no or unsure?” participants often would overestimate their level of happiness. Because blog posts and Tweets are unsolicited content, this effect is much less likely to occur.
However, there are problems with using the software. Bloggers and Twitter users tend to be younger than the average population, for example.
A bigger issue is the limits of the software itself. The sentence “I am not happy” registers as a happy statement.
