Twitter mood API

Question

Is there a twitter API that returns mood? I have seen some websites that return the mood of a specific tweet and I want to be able to use an API (or to use some open source code) to get the mood of a tweet, does anyone know if there is an easy way to do that or if there is an already built API for this?

Answer 1

Here is a list of APIs/websites that implement sentiment analysis on text.

Hope it helps. :)

https://sites.google.com/site/twittersentimenthelp/other-resources

Answer 2

The link in the accepted answer is no longer working and the rest of the links in the other answers are all pretty useless. The best Twitter sentiment project and accompanying API is this one from Stanford. It's better then the defunct tweetfeel.

Answer 3

I did a project involving Sentiment Analysis of Tweets. I used a machine-learning sentiment analyzer called 'LingPipe'. Unfortunately, it requires you to design your program in Java (I know you tagged your question under PHP), and to make a fairly decent training set from your raw data (very tedious work). However, if you do have Java skills, and time, then it's fairly easy to use, and extremely speedy.

Answer 4

There's a new free API from TweetFeel that uses advanced linguistic processing to determine sentiment (mood). Take a look at http://www.webservius.com/corp/docs/tweetfeel_sentiment.htm

Answer 5

You'd probably want to use another API to do sentiment analysis on the tweets you want to extract moods from. Open Amplify is probably a good one to start with http://community.openamplify.com/content/docs.aspx/.

Answer 6

I don't think I've seen any Twitter tweets mood API around, but someone made it possible through the use of Twitter API. http://i8news.uterm.org/mood/twitter-mood-reader/

The logic was to read through the tweets and by calculating the emotional words (such as "wow", "happy", "!", "darn"), you get the overall mood of that tweet.

Answer 7

Most of the websites you see doing that are using a general technique called sentiment extraction. There are many possible implementations, but a simple (somewhat naive) one is to make a list of "bad" words ("shouldn't", "angry", "evil", "allergic") and a list of "good" words ("happiness", "joyful", "smiling", "wonderful"). Assign a "mood score" to each word (or just -1 to bad words and +1 to good words). The total score is the "mood'.

This is a naive approach because many words require context to understand. (For example, a Bostonian describing something as "wicked good" is making a doubly positive remark, not canceling out the following word.) Still, it may be suitable for your purposes.