Twitter text mining for sentiment analysis of COP15

Posted on Updated on


In the project Responsible Business in the Blogosphere between Copenhagen Business School and Technical University of Denmark we will try to use text mining on blogs for computerized analysis of so-called corporate social responsibility (CSR). Twitter provides a convenient interface for downloading the tweets on its site and as a start on the project I tried to see how this microblogging service could be used in sentiment and topic analyses. A prototype is now running from the Web server, and even a rudimentary sentiment analysis seems to find that for example the company Novo Nordisk has more positive tweets than for example Pfizer has.

As of ultimo November 2009 I have put a sentiment analysis of tweets relating to (United Nations Climate Change Conference (COP15) on the main page. It shows a “barometer” reflecting the ratio of positive and negative tweets. The negative tweets may for example be from twitters pessimistic about how we can solve negative consequences of human impact or they may be from twitters hostile to the COP15 meeting. Positive tweets may for example be from twitters excited about going to Copenhagen or from twitters expressing hope.

A positive example (slightly edited):

Barack Obama to attend Copenhagen climate summit. Ah, that’s better news – hope he brings good US stance

A negative one (also slightly edited):

I Hate Al Gore ?? Global Warming Fraud Exposed as Copenhagen Approaches

It will be interesting to see the development of sentiment related to COP15, especially during the conference days 7 December to 18 December 2009.


3 thoughts on “Twitter text mining for sentiment analysis of COP15

    cyhex said:
    July 13, 2011 at 9:50 am

    Take a look at Twitter sentiment analysis tool , its written in python and uses Naive Bayes classifier with semi-supervised machine learning

    Anonymous said:
    July 13, 2011 at 10:11 am

    @cyhex,Thanks cyhex. I have seen that interesting web service before. Thanks for reminding me.I wonder what the extra "semi" besides the Naive Bayes classifier is?

    cyhex said:
    July 13, 2011 at 11:01 am

    semi-supervised machine learning is a method that lets you train data with small amount of trained and large amount of untrained data.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s