Sunday, December 8, 2013

What is twitter buzzing about Dhoom3?

Hi Friends,
    In the past, you might have already seen a lot of movie reviews written by plenty of critiques out there on the internet. No matter how genuine a critique is, as long as he is human, there is always a tendency to introduce 'bias' factor into his review. It is very rare that you get 'actual' public opinion from a single reviewer!
  In this context, I gave a small practical attempt to pull out 'public pulse' of what people are buzzing on twitter.
As we all know dhoom 3 is set to release in a couple of weeks, I pulled out twitter tweets tagged with #Dhoom3 (systematic sampling) from last 20 days and performed some text mining.
Here are the findings
Lets delve into what are common buzz words around #dhoom3. Here's a wordcloud of what public is tweeting about the movie. (bigger word size implies more frequently used word in tweets)

 Now, I tried to make a few associations between words and here is list of words that are commonly buzzed about around frequent terms

Malang -> associated frequently with (dum,ishq,song , soexcited)
[Looks like fans are indeed exited about this song!!]
Aamir -> associated frequently with (khan, 5crore, underwent, medication,nobody)
[Ohh.. Looks like song budget and his medication are trending with aamir]
Katrina-> associated frequently with ( kaif, kaif’s, acrobats, gain , loss,  training )
[Nothing unusual :( ]
Kamli ->  associated frequently with (  katrinakaif, showbiz, sizzles , promo )

 Finally, An attempt to group related words into distinct buckets brought down the following clusters of words that people commonly tweeted!!
[Top 8]
cluster 1: kaif katrina dhoom3 dhoom kamli
cluster 2: aamirkhan dhoom3 dhoom movie days
cluster 3: song dhoom3 malang promo dhoom
cluster 4: aamir dhoom3 dhoom khan indeals
cluster 5: dhoom3 hot looks getting better
cluster 6: 20th big cant december dhoom3
cluster 7: malang dhoom3 dum ishq song
cluster 8: dhoom3 malang back dhoom bollywood

What next??  wait for actual public sentiment coming up after movie is released!

[And for geeks out there, science behind the scenes: R-statistical language, java & twitter4j api, Datamining, Text mining, Analytics ]


Friday, December 6, 2013

Text mining public tweets of #fabindia

Hi Guys
    Today I had a chance to crawl over some  public tweets tagged with #fabindia as a part of my academic project to figure out sentiment analysis of its customers.
Look at the word cloud that I built to give you a heads up of what people are buzzing about fab india.



Although no. of tweets out there are pretty small in number, this analysis gives a headway to what kind of potential public sentiment be extracted from twitter.
And yes hand on!! wait for much more interesting text mining results that are yet to come over in coming days! (#bigdata, #datamining #R )


Cheers!!