Movies: Art or Business?
Photo from Unsplash
Is movie making primarily an art or a business? Idealistically, it's a medium to expand our mind or imagination, evoke emotions, for storytelling, for demonstrating cinematography.
However, does the movie content matter if it makes a profit, if it attracts a crowd to the movie theatres?
At the end of the day, it is still a business, and if audiences go to watch movies that are not "artistic", or if "good" movies are not profitable, unfortunately, businesses will follow the money.
So is this what is happening in real life?
Some questions that I had in mind when approaching the project were as follows:
- What is the distribution of film ratings for the top 100 grossing films each year?
- Are there any noticeable trends over time on the relationship between the world's top grossing films and its film rating? (i.e. have film ratings of the top grossing movies gone down or up?)
- Are audiences' movie tastes or motivations for going to a movie theatre changing?
I planned to see if I could see these relationships by digging up the top grossing films and its general finances, along with its corresponding RT critic score and user score, to visualizing these factors.
Data scraping & analysis
Using Scrapy, I scraped the top 100 grossing movies each year from 1989 to 2019 from "the-numbers.com". Attributes included:
|Worldwide Box Office numbers|
|Domestic Box Office numbers|
|Production Budget (if available)|
I then also retrieved the corresponding RottenTomatoes ("RT") rating for each of these top 100 grossing films for each year. RottenTomatoes is an online aggregator of movie reviews from critics, and reviews from the general audience:
|Number of critic reviews|
|Number of user reviews|
The RT Tomatometer ("Critic scores") is the aggregated scores given by critics on a certain movie. The RT audience meter ("User scores"), is the score given by users/audiences on a certain movie. (These ratings will come into play in conducting the data analysis, shown below.)
In conducting the analysis, comparisons were limited to comparing two decades: (1) movies from 2000 to 2009, and (2) movies from 2010-2019.
RottenTomatoes Critic Reviews vs. User Reviews
Firstly, to get a lay of the land. What is the general distribution of how critics rated movies versus how users rated movies?
Plot 1 and Plot 2 below show a summary of the distribution of critic scores versus user scores.
We can see that from 2000 -2009, there was a larger spread of points from the linear regression line, which suggests that there is more deviation in Critic vs. User scores in this decade.
Plot 1: RT Critic Score v. User Score, 2000-2009
Comparing the linear regression line of movies within the "Fresh " category (blue line), we can also see that Critics and Users scores of movies in this category were much more aligned from 2010-2019.
Plot 2: RT Critic Score v. User Score, 2010-2019
Critic-User Difference versus its performance in theatres
In Plot 3 and Plot 4 below, the difference between the Critic scores and User scores (called "Critic-User Difference") is plotted against its Worldwide Box Office Profit (Worldwide Box Office minus Production Budget).
Where a Critic-User Difference is negative (e.g. -60), it means the User score was 60 points higher than the Critic score.
Theoretically, where the Critic-User Difference is 0 (i.e. Critics and Users are in agreement), we should be seeing movies all along the y axis (i.e. a good movie (that critics and users agree are good movies (“Certified Fresh”/“Fresh”)) performed well and earned, and a bad movie (that both critics and users say are bad (“Rotten”)) did not perform well.
Plot 3: RT Critic-User Difference v. Worldwide Profit, 2000-2009
In Plot 3, we can even see that there are quite a few outliers on the right hand side where “Certified Fresh” movies are raking in quite a good profit.
In contrast to Plot 3, when looking at Plot 4 below, the linear regression curve is downward sloping as it moves along the x-axis (“Critic-User Difference”) for movies from 2010 – 2019.
There can be two factors contributing to this downward slope:
- Change in production budget: Studios are spending more on the production budget for films that critics subsequently also rated as good films (“Certified Fresh”/”Fresh”) but it has also reduced the profitability of such films, or studios are spending less on films that critics are rating as bad films (“Rotten”), but its profitability has increased because of the lower associated costs.
- Change in box office: More people are going to watch films that Critics are calling “Rotten”, or less people are going to watch films that Critics are calling “Fresh”.
In this instance, there are more movie titles in 2010-2019 which are seeing greater profit who are on the left hand side of the graph.
Plot 4: RT Critic-User Difference v. Worldwide Profit, 2010-2019
Critic-Users Difference versus its performance in theatres relative to its costs
Lastly, we will look at the Critic-Users Difference in relation to its Worldwide Box Office/Production Budget ratio.
Worldwide Box Office to Production Budget ratio (“Earning Ratio”) is calculated by taking the film’s worldwide box office and dividing it by its production budget. Here, we would get a measure of how much a film is earning relative to its costs.
Plot 5: RT Critic-User Difference vs. Box Office/Production Budget Ratio, 2000-2009
In Plot 5, we will see there is a positive relationship between a film’s rating and its Earning Ratio from 2000 to 2009. A high Critic score for the movie corresponds to a higher Earning Ratio.
It could mean that the audiences went to watch good films, and these were also movies rated as “Certified Fresh”/“Fresh” by Critics.
In Plot 6, we can see that from 2010-2019, the relationship between the two is not as high.
Plot 6: RT Critic-User Difference vs. Box Office/Production Budget Ratio, 2010-2019
Conclusions & Further thoughts
The landscape for movies seems to have changed from 2000 to 2019. Though it does look like the audience behavior seem to have changed between 2000-2009 and 2010-2019, I would want to conduct further analysis before coming to a definitive conclusion.
Given more time, I would like to perform a further deep dive to obtain a deeper understanding of the factors driving the numbers behind the movies’ performance in theatre.
This would include a sentiment analysis on reviews of movie goers on specific films to understand what motivated them to go watch the movie, and how much marketing was conducted to promote the film. It would also be interesting to do a breakdown and analysis by movie genre or the source of the screenplay (e.g. based on a book, original screenplay, based on comic) to gain further insights on what factors are most effective in bringing audiences to the movie theatre.
Hope you enjoyed the read!
The skills I demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.