Tuesday, July 21, 2015

Movie sentiment analysis - Part 1

Prologue :

Frankly, I am very new to this data mining realm. At first it felt like I was going somewhere just after learning the language R. Like my father said, everyone has a bow and an arrow, archer should be Arjun to hit the target.

Later, it was fear of being lost. It doesn't make sense. We actually relearn and redo stuff when we dive into it. So, I had decided to take up analysis and comparison to start with.

Synopsis:

This project deals with sentiment analysis of live twitter feed to compare two movies that are fighting for the throne harder than GOT characters - Bahubali and Bhajrangi Bhaijaan. I collect the data and polish it in such a way that it remains true to the source removing unwanted garbage of NLP. Then it creates a word cloud giving us a picture of what we are dealing with in content. We apply a simple rating system to keep a score of performance of the picture. It lasts for roughly three months and predictions will be supplied every day. The average predictions for day forms monthly prediction.

Objective:

To predict the performance of each movie in terms of public opinion and deriving a relation between public opinion and box office collections.

Introduction :

Bahubali is an Indian quasi-historical fiction film directed by Mr. S.S. Rajamouli with a star cast consisting of 7 main characters - Prabhas as Bahubali, Rana Daggubati as BhallalaDeva, Anushka as DevaSena, Tamanna as Avantika, SatyaRaj as Kattappa, RamyaKrishna as Sivagami and Nassar as BijjalaDeva.

Bhajrangi Bhaijaan is an Indian contemporary movie directed by Mr. Kabir Khan starring Salman Khan in the lead role and Kareena Kapoor as his love interest in the film.

The question upon which everyone is pondering over the internet since last Friday is which of these films would be biggest hit at the box office. 
At one side, we have "Bhai of India" or quite clearly Robert Downey Jr of India, Mr. Salman Khan who is known for massive following nowhere less than a cult in India and overseas.
On the other side we have "Director of the Century" or some might call "Peter Jackson of India" for his experimental and quite spectacular visuals embedded in a delicate story that is larger than life.

The polarity of people towards one film or the other is actually not worth investigating if pessimistically attributed to mob mentality. It should also be not taken too seriously for it practically pits couple of hunks to a battlefield (no pun intended). It is purely academic concern to test the winning strategy.
Bahubali relies on story, visuals to depict medieval India and success of the director as opposed to Bhajrangi Bhaijaan which explores story, sensitive case of Indo-Pak relationship and puts heavy reliance on Salman's fans.
Both stories have one thing in common for sure. The screenplay is by same person, K. Vijayendra Prasad. 

My concern with the film hit me when the chase became interesting. Both films were a week apart in release dates with Bahubali going first. Some started complaining about portrayal of women while some went to lengths to disown an Indian film citing it as a south Indian movie. This started similar and stronger opposition in North Indian community itself, ironically where people taunted Salman's acting failure and his lack of being in a good story, even though he had a great box office record. When the film Bhajrangi Bhaijaan released, people went nuts for they got an unexpected surprise of "Bhai" being in a film with great story. The chase reversed as people in South India praised Bhai and showered him with compliments, again ironically. 

The level of people's commitment is augmented to the fact that all the bars have been raised. Naysayers keep waiting for any copied material from Hollywood ignoring the effort that goes into film making even if it be a copy. Intellectuals try to grab the corpus of the film disintegrating in qualitative and quantitative subject-wise analysis. Common folk measure entertainment they've been starving for, effectively since Happy New Year and Dil Dhadakne Do.

The fluid paradigm shifting behavior of people with each passing moment is precious. We can't get public opinion as fluid as this until next election of a large democracy. Well India had its turn, so yeah, U.S.America.

My argument that it is more precious as elections might say quantitatively about who might win but this analysis is capable of giving us an insight into what people want more - a hyped movie or a sleeper hit, a medieval pride or a contemporary achievement , may be a sense of Indian pride or pure entertainment?

It is actually unfair to deduce the conclusion from just one fight, but this fight is no less than fight between Batman and Superman. Bhai is Indian Batman as in Kick (ironically south Indian movie copy but whatever), and Bahubali literally means one with strong arms (KJo compared it with Mughal-E-Azam for some reason?). I mean, we all know Batman wins naturally but you get the point. It's not about who/what, it's how. How much public opinion account in box office collections? Will it be like Avatar which was visual marvel but had recycled plots or like Terminator 2 that was great in all departments? Will it be like Bad Boys with OMG results or like After Earth which is film equivalent of diarrhea? 

Keep following at https://github.com/Horopter/MovieSentimentAnalysis for more.

AstalaVista Bros!

No comments:

Post a Comment