1 of 16

Movie Data

Group 18: Jack Kraska, Paula Andrei Chiang, Megha Rajan, Lauryn Larzo, Braydon Stern

Case 3

2 of 16

Meet Jessica

2

Occupation: Business Data Analyst at WPC+

Context: WPC+ is a new streaming platform that is currently in development

Objectives:

Identify a set of blockbuster movies that people will watch based on IMDb scores
Examine what variables impact IMDb scores the most as WPC+ has plans to create its own movie content in the future
Report back with Recommendations

3 of 16

Data Variables

3

Duration
Budget
Color (Black/White or Color)
Director, Actor 1, Actor 2, Actor 3, Cast Total, Movie Facebook Likes
Number of Critics, Number of Users that voted for the movie, Number of Users that reviewed the movie
Year
Face Number in Poster
Aspect Ratio

IMDb score

4 of 16

Data Cleaning

4468 rows of data

Checked for missing variables via Columns Viewer

found 599 rows of data missing for ‘Gross’ so we excluded it

Test for any abnormalities in the distribution
Did a multivariate analysis and eliminated outliers via Mahalanobis Distances

4

5 of 16

Questions and Data Analysis

5

6 of 16

Is there a correlation between movie duration and IMDb score? (Black & White vs. Color)

Filtered by

color of the movie

6

Paula

Graph on the left

-Null hypothesis: duration by black/white and imdb score are not significantly correlated

-alternate hypothesis: duration by black/white and imdb score are significantly correlated

Graph on the right

-Null hypothesis: duration by color and imdb score are not significantly correlated

-alternate hypothesis: duration by color and imdb score are significantly correlated

It is important to note the distribution of the color of the movie as there is significantly more color movies than black and white in the data set hence the difference in # of data points on the graphs
Based on the r-square in both graphs, imdb score does not have the greatest correlation with duration regardless of the color of the movie.
The black and white graph on the left you would fail to reject the null hyp because of the p-value of 0.68
The color graph on the right you would reject the null hyp because of the p-value is less than 0.05 suggesting a correlation between duration by color and imdb score
The longer the movie (Colored), the greater the impact on IMDB score

7 of 16

Does the number of critic reviews affect the IMDb Score? (Black & White vs. Color)

7

Let’s start by looking at the analysis for Black & White movies

Next, let’s look at Color movies

8 of 16

Does the number of critic reviews affect the IMDb Score? (Black & White vs. Color) Cont.

8

Key Takeaways:

Although we only have 60 points of black & white movie data compared to the 1840 points for colored movies, the r² fit line for B&W movies is twice as accurate as the fit line for colored movies.

B&W movies are more likely to have a higher IMDb score compared to colored movies with a similar number of critic reviews.

9 of 16

Does the number of Movie Facebook likes have a correlation with the IMDb Score?

Filtered to include 2004 to 2016

9

10 of 16

Which years/movies have the best IMDb score?

10

11 of 16

How do certain variables affect the IMDb score?

11

12 of 16

12

13 of 16

13

14 of 16

Jessica’s Recommendations/

Conclusion

If WPC+ wants to generate its own content, the budget and the duration of the movie are important factors to consider with regards to the success of a movie.

duration by color movies is more important than black and white when it comes to a higher IMDb score
number of critics and movie facebook likes does correlate with IMDb score regardless of the color of movie

Based on IMDb score, it would be good to include movies from the following years: 1933, 1959, 1964, 1969, and 1973

14

15 of 16

Citations

15

16 of 16

Thank You!

Questions?

16