Movie Trend in Indonesia

Intro

It is important to know how the development of movie lovers for film practitioners and also the entertainment world today.

Through Google Trends, we can now see how people in Indonesia tend to look for info related to the films they like.

However, there are so many types of films that it might be difficult for us to find out which film genre has a big influence on films in Indonesia.

Therefore, in this aricle, it will be shown how the influence of the film genre with the Movie that develops in Indonesia based on search interest.

Limitation

• The independent variable or the variable that is affected is the movie data
• The dependent variable or the influencing variable is the variable ‘drama’, ‘comedy’. ‘horror’, and ‘action’

With data taken from Google Trends, it can be seen that the development of film trends in Indonesia is as follows:

Trend Movie (past 12 months) in Indonesia

If you look at the average development of movies for the horror genre with drama, there is not too much difference, meanwhile comedy and action also have a similar pattern.

The visualization above is then converted into data like the following.

Model

The model that will be used is multiple regression analysis.

Call:
lm(formula = regresi$movie ~ regresi$drama + regresi$horror + 
    regresi$comedy + regresi$action)

Residuals:
     Min       1Q   Median       3Q      Max 
-14.2032  -3.6661   0.2063   4.3156   9.4775 

Coefficients:
               Estimate Std. Error t value Pr(>|t|)    
(Intercept)     27.4386     5.9241   4.632 2.89e-05 ***
regresi$drama    0.9758     0.3796   2.570   0.0134 *  
regresi$horror   1.3193     0.5266   2.505   0.0158 *  
regresi$comedy   3.1972     1.2053   2.653   0.0109 *  
regresi$action   2.0822     1.3330   1.562   0.1250    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.001 on 47 degrees of freedom
Multiple R-squared:  0.5926,    Adjusted R-squared:  0.558 
F-statistic: 17.09 on 4 and 47 DF,  p-value: 1.02e-08

The regression equation model is obtained as follows:

y(movie) = 27.44 + 0,98(drama) + 1.32(horror) + 3.2(comedy) + 2.1(action) + error

Hypothesis:

• h0 : There is no influence between the independent variable and the dependent variable
• h1 : There is at least one effect on the movie variable

taking the error rate of 0.05, therefore, p-value<0.05. Thus, the decision is to reject h0, meaning that there is a simultaneous influence on the movie variable.

---

If viewed partially, we can see that the variable ‘action’ has a p-value greater than 0.05. Thus, the variables that affect the movie variable are the ‘drama’ variable, the ‘horror’ variable, and the ‘comedy’ variable.

Therefore, the model was updated again and the following is the result of the latest model.

Call:
lm(formula = regresi$movie ~ regresi$drama + regresi$horror + 
    regresi$comedy)

Residuals:
     Min       1Q   Median       3Q      Max 
-14.0930  -3.3048   0.7913   3.8233   9.7809 

Coefficients:
               Estimate Std. Error t value Pr(>|t|)    
(Intercept)     31.6126     5.3660   5.891 3.67e-07 ***
regresi$drama    1.0006     0.3849   2.599  0.01237 *  
regresi$horror   1.4525     0.5273   2.754  0.00828 ** 
regresi$comedy   3.3160     1.2208   2.716  0.00915 ** 
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.09 on 48 degrees of freedom
Multiple R-squared:  0.5715,    Adjusted R-squared:  0.5447 
F-statistic: 21.34 on 3 and 48 DF,  p-value: 6.327e-09

Model II

y(movie) = 31.61 + 1(drama) + 1.45(horror) + 3.32(comedy) + e

Based on the results of the model above, each variable has an effect either partially or simultaneously on the movie variable.

Plot Model

Classic Assumption Test

Multicollinearity Test

 regresi$drama regresi$horror regresi$comedy 
      1.324169       1.789534       1.479751 

Based on the results of VIF (Variance Inflation Factor), the VIF value obtained by each variable is < 10. Therefore, it can be concluded that the Non-Multicollinearity assumption is fulfilled.

Normality test

    Jarque Bera Test

data:  model$residuals
X-squared = 1.9807, df = 2, p-value = 0.3714

With the error rate taken is 5%, then based on p-value> 5%. Therefore, the model satisfies the assumption of normality over the residuals.

Autocorrelation Test

    Breusch-Godfrey test for serial correlation of order up to 1

data:  model
LM test = 7.506, df = 1, p-value = 0.00615

With the error rate taken is 5%, then based on p-value> 5%. Therefore, the model satisfies the assumption of autocorrelation over residue

Heteroscedasticity Test

    studentized Breusch-Pagan test

data:  model
BP = 0.6198, df = 3, p-value = 0.8919

because the p-value is greater than 0.05 then the non-heteroscedasticity assumption is fulfilled.

Interpretation

y(movie) = 31.61 + 1(drama) + 1.45(horror) + 3.32(comedy) + e

• The ‘Drama’ genre regression coefficient of 1 states that every one point increase in the ‘Drama’ genre will cause an increase in the movie by 1 point assuming the other variables are constant.
• The ‘Horror’ genre regression coefficient of 1.45 states that every one point increase in the ‘Horror’ genre will cause an increase in the movie by 1.45 points assuming the other variables are constant.
• The ‘Comedy’ genre regression coefficient of 3.32 states that every one point increase in the ‘Comedy’ genre will cause an increase in the movie by 3.32 points assuming the other variables are constant.
• If there is no increase in each variable, the movie variable will continue to increase by 31.61 points.

---

Thus it can be concluded that the variables ‘Drama’, ‘Horror’ and ‘Comedy’ have an influence on the variable ‘Movie’ in Indonesia based on search interest with the following model:

y(movie) = 31.61 + 1(drama) + 1.45(horror) + 3.32(comedy) + e

the percentage of the influence of the model is 54.47%, the variables can represent this influence and the rest are influenced by other variables.

---