Twitfluence application basic technical specifications

icon

Twitfluence is a registered Twitter application for calculating the "weight" and influence of your Twitter account, and is already accessible for beta users. It uses read and write permissions on your Twitter account. Perhaps it will be upgraded with write permissions one day too, so it will be able to post the results to your timeline on request, but for now, read permission is all it needs. The original idea was to have it done without Twitter authentication (by simply entering a user name), but then you're not able to access mentions and retweets, which are obviously a big thing in measuring someone's Twitter influence.

The data

At this stage, the Twitfluence calculation uses and stores the following information for it's calculation. None of your personal data, your authentication info and your tweets are stored, so you need to approve it each time you do the calculation. Here is the full list of the parameters captured:

  • You
    • How many people you follow
    • How many people follow you
    • How many days you are on Twitter
    • How many Tweets you've made
  • One month of tweets
    • How many Tweets you've made
    • How many mentions and replies you’ve made
    • How many retweets you’ve made
    • How many times you were mentioned and replied to
    • How many times you were retweeted
    • What was the reach of you being mentions and replied to
    • What was the reach of your tweets retweeted
  • Lists
    • How many list you appear on
    • How many people these list follow
    • How many people follow these lists
    • How many people follow the people on whose lists you appear
    • How many lists you own
    • How many people follow your lists
  • Other
    • Your screen name
    • Your profile picture
    • Your description
    • Your timezone for comparison to nearby users

API calls

Twitfluence calculation makes 6 API calls to capture the data mentioned. Here is the full list of the requests:

  • http://api.twitter.com/1/statuses/user_timeline.xml: your tweets
  • http://api.twitter.com/1/statuses/retweeted_by_me.xml: you retweets
  • http://api.twitter.com/1/statuses/mentions.xml: you being mentioned
  • http://api.twitter.com/1/statuses/retweets_of_me.xml: you being retweeted
  • http://api.twitter.com/1/screenname/lists.xml: list you own
  • http://api.twitter.com/1/screenname/lists/memberships.xml: list you appear on

Retweets

Twitter allows different clients (web, mobile clients,…), and some of them make retweets in a technically different way that the core Twitter web client. Therefore retweets are scattered inside retweets and mentions, but the Twitfluence application is smart enough to recognize them. The following tweets are consider retweets, and the calculation uses retweets in a different way than mentions and replies:

  • actual retweets
  • mentions that contain the syntax "RT @"
  • mentions that contain the syntax "via @"

Reach

There are a few parameters inside the calculation that use the "reach" of tweets (retweets, mentions, lists). You should know that this is not the actual real reach (potential unique users) of your tweets without duplicates, but a simple addition of the followers for the specific case. Therefore the score from reach may not be perfectly correct, but at this point it will have to do.

Links

Clicking on links published on Twitter is currently not a part of the calculation. Measuring clicks can be a bit difficult, specially if the links point directly to the website (compared to using URL shorteners, which offer analysis of clicks). Hopefully I will be able to use links in Twitfluence as well someday, in a way that's proper and objective for most users, but for now they'll be ignored.

This is the basic technical specification of the Twitfluence application. In case your curious about the mathematical algorithm behind it and how the calculation works, you can visit the following post, but otherwise, feel free to leave a comment or suggestion. I want to make the calculation as correct as possible, so any feedback of yours would really be appreciated.

Like what you read? Now tell the world!
Vote on Hacker News

A few more things you might find interesting:


Comment
written 1.8.2010 12:53 CET on chronolog
1151 views   •   1 like   •   Like   •   
date
date
date
date

Connect with :