Monday, June 15, 2009

What the Little Bird Told Me About You: Three Twitter Apps for Psych Analysis


Tomorrow morning, social media and marketing researcher Dan Zarrella is debuting a new way to see into the minds of Twitter users by analyzing their most recent 1,000 tweets.

TweetPsych uses two linguistic analysis methods to build a psychological profile of a person based on the content of their tweets. It compares the content of a user's tweets to a baseline reading Zarella built by analyzing over 1.5 million random tweets and shows the areas where that user stands out. It also reminded us of two other fascinating apps that show how long a user has been on Twitter and with whom they hold most of their @reply conversations. Being socially minded journalists, we've made bookmarklets for all three services.


Zarrella wrote in an email tonight that he used RID (Regressive Imagery Dictionary) and LIWC (Linguistic Inquiry and Word Count) to parse the data. RID is a text analysis tool composed of more than 3,000 words from 43 categories of cognition and emotion. LIWC is a text analysis software program that calculates the degree to which people use different categories of words in emails, speeches, poems, or transcribed daily speech. The program considers positive and negative emotion words, self-references, and words that refer to sex, eating, or religion.

Profiles with updates that are protected cannot be analyzed by TweetPsych.

Let's take a look inside the mind of a few Twitter users. Most of the social media elite tend to have fairly impersonal tweets; hence, their TweetPsych profiles are relatively homogeneous catalogs of upward mobility, obsession with professional affairs, and moral imperativism. Here's a profile of a photographer/mother/homemaker/blogger in Georgia:

In marked contrast, here's a 20-something, male entrepreneur in Virginia:

We thought TweetPsych was so nifty that we, a.k.a. Marshall Kirkpatrick, made a bookmarklet. Drag the text TweetPsych into your browser's bookmar bar, visit a Twitter profile, then click the bookmarket to see an analysis of a Twitter user's profile.

Other tweet-analyzing apps we love are Mailana, which shows Twitter conversations and links between different users within and beyond a given user's network, and WhenDidYouJoinTwitter, which shows the date a user joined Twitter (or the date the user implemented the most recent iteration of his username). The WhenDidYouJoinTwitter bookmarklet is also available at that link.

Here's Mailana at work:

This app is particularly good at showing the hubs or connectors in your network, and can also be useful for making new connections with other users. You can use the Mailana bookmarklet on any Twitter profile.

Best of all, try out Marshall's 3-in-1 bookmarklet LittleBirdie to see what each of these apps finds from the Twitter users you love best. Simply drag the text link into your browser's bookmarks bar, visit a Twitter profile, and start analyzing/judging the heck out or everyone you do (and don't) know.