Getting valuable insights from TMDb movies dataset.

Charles Maina Mwangi — Wed, 08 Jun 2022 09:46:24 +0000

I have had an amazing opportunity to participate in a Nanodegree in Udacity in the past month and I want to share my interesting journey. I am an aspiring data analyst and in this blog, I will share my first project for investigating the TMDb movie dataset.

Project: Investigate the Movies Dataset

Introduction
Data Wrangling
Exploratory Data Analysis
Conclusions

Introduction

This dataset is dowanloaded from kaggle.com. It is the Movies dataset. It contains information about over 10000 movies that are collected from The Movie Database (TMDb). The dataset includes the movie title, directors, cast, among others. Some of the columns like cast, keywords, and genres contains multiple values and separated by pipe (|) characters.
The following are the questions I ask myself for this analysis.

Research Question 1 (Which movie genre is the most popular?)

Research Question 2 (Who is the most common director for movies?)

Research Question 3 (Which are the most common keywords?)

Research Question 4 (Which production companies produce the most movies?)

Research Question 5 (Which is the most common tagline?)

# Use this cell to set up import statements for all of the packages that you
#   plan to use.
# Importing required libraries
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from collections import Counter
%matplotlib inline

Data Wrangling

This step will involve loading the dataset, cleaning the dataset and trimming it.

General Properties

# Load your data and print out a few lines. Perform operations to inspect data
#   types and look for instances of missing or possibly errant data.
movies_df = pd.read_csv("tmdb-movies.csv")
movies_df.head(5)

# Check number of rows and columns
movies_df.shape

(10866, 21)

# Describe the data
movies_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 10866 entries, 0 to 10865
Data columns (total 21 columns):
 #   Column                Non-Null Count  Dtype  
---  ------                --------------  -----  
 0   id                    10866 non-null  int64  
 1   imdb_id               10856 non-null  object 
 2   popularity            10866 non-null  float64
 3   budget                10866 non-null  int64  
 4   revenue               10866 non-null  int64  
 5   original_title        10866 non-null  object 
 6   cast                  10790 non-null  object 
 7   homepage              2936 non-null   object 
 8   director              10822 non-null  object 
 9   tagline               8042 non-null   object 
 10  keywords              9373 non-null   object 
 11  overview              10862 non-null  object 
 12  runtime               10866 non-null  int64  
 13  genres                10843 non-null  object 
 14  production_companies  9836 non-null   object 
 15  release_date          10866 non-null  object 
 16  vote_count            10866 non-null  int64  
 17  vote_average          10866 non-null  float64
 18  release_year          10866 non-null  int64  
 19  budget_adj            10866 non-null  float64
 20  revenue_adj           10866 non-null  float64
dtypes: float64(4), int64(6), object(11)
memory usage: 1.7+ MB

# Check dataset datatypes
movies_df.dtypes

id                        int64
imdb_id                  object
popularity              float64
budget                    int64
revenue                   int64
original_title           object
cast                     object
homepage                 object
director                 object
tagline                  object
keywords                 object
overview                 object
runtime                   int64
genres                   object
production_companies     object
release_date             object
vote_count                int64
vote_average            float64
release_year              int64
budget_adj              float64
revenue_adj             float64
dtype: object

Cleaning the data

Here, we first drop all duplicates present in the dataset.

The next step involves checking the number of null values present in each column. After determining the null values, we then drop them all and check the dataset for any other null or missing values.

# Drop all duplicate values. 
movies_df.drop_duplicates(inplace=True)

# Check all null and missing values
movies_df.isnull().sum()

id                         0
imdb_id                   10
popularity                 0
budget                     0
revenue                    0
original_title             0
cast                      76
homepage                7929
director                  44
tagline                 2824
keywords                1493
overview                   4
runtime                    0
genres                    23
production_companies    1030
release_date               0
vote_count                 0
vote_average               0
release_year               0
budget_adj                 0
revenue_adj                0
dtype: int64

# Remove and drop all null values and missing values
movies_df = movies_df.dropna(subset=['imdb_id','cast','homepage','director','tagline','keywords','overview','genres','production_companies'])

# Check again to confirm all the missing and null values have been droppped. 
movies_df.isnull().sum()

id                      0
imdb_id                 0
popularity              0
budget                  0
revenue                 0
original_title          0
cast                    0
homepage                0
director                0
tagline                 0
keywords                0
overview                0
runtime                 0
genres                  0
production_companies    0
release_date            0
vote_count              0
vote_average            0
release_year            0
budget_adj              0
revenue_adj             0
dtype: int64

Exploratory Data Analysis

Research Question 1 (Which movie genre is the most popular?)

# Function to count the top 15 most common in the dataset
def counting(param):
    counter = Counter()
    for i in param:
        counter[i] += 1
    top15_most_common = counter.most_common(15)
    return top15_most_common

# Continue to explore the data to address your additional research
#   questions. Add more headers as needed if you have more questions to
#   investigate.
genres = "|".join(movies_df['genres']).split("|")

top15_most_common_genres = counting(genres)


genres_df = pd.DataFrame (top15_most_common_genres, columns = ['genres','Genres Count'])
genres_df.sort_values(ascending=True, by='Genres Count', inplace=True)
ax = genres_df.plot.barh(x='genres',color=["#FF69B4","#F0F8FF"],width=0.7,legend=False, figsize=(10, 10), title="A bar plot to find the most common genre of the movies")
ax.set_ylabel("Genres")
ax.set_xlabel("Genres counts")

Text(0.5, 0, 'Genres counts')

This bar chart is used to analyze the most common genrein this movie dataset. The analysis is done for the top 15 most common genres with their number of counts in the dataset. It is clear that most of the movies produced directed are in the Drama Genre type. This is followed by Comedy, Thriller, Action and Adventure movies. It measn that these are the most watched movies.

Research Question 2 (Who is the most common director for movies?)

director = "|".join(movies_df['director']).split("|")

top15_most_common_directors = counting(director) 

director_df = pd.DataFrame (top15_most_common_directors, columns = ['director','Directors'])
director_df.sort_values(ascending=False, by='Directors', inplace=True)
ax = director_df.plot.bar(x='director',color=["#746AB0"],width=0.7,legend=False, figsize=(10, 10), title="A bar plot to find the director who produced the most movies. ")
ax.set(xlabel="Directors", ylabel="Directors counts")

[Text(0.5, 0, 'Directors'), Text(0, 0.5, 'Directors counts')]

In research question two, I went further to investigate which director had produced the most movies from this dataset. I did that by finding the most common director by ther counts. I picked the top 15 most common to make the plot as shown above. I found out that John Carpenter was the most common director. Hence, John Carpenter was the director who had produced the most moves.

Research Question 3 (Which are the most common keywords?)

keywords = "|".join(movies_df['keywords']).split("|")
keywords

top15_most_common_keywords = counting(keywords) 

keywords_df = pd.DataFrame (top15_most_common_keywords, columns = ['keywords','Keywords'])
keywords_df.sort_values(ascending=True, by='Keywords', inplace=True)
keywords_df.plot.barh(x='keywords',color=["#FFCE30"],width=0.7,legend=False, figsize=(10, 10),xlabel="Keywords", ylabel="keyword counts", title="A bar plot to find the most common used keyword in naming movies.")

<AxesSubplot:title={'center':'A bar plot to find the most common used keyword in naming movies.'}, ylabel='Keywords'>

I went further to find out which keyword was the most used in the movies. I was able to find this by analyzing the dataset for the most common top 15 keywords used by their counts. I found out that based on the novel was the most used and most common keyword.

Research Question 4 (Which production companies produce the most movies?)

movies_df['production_companies'].value_counts()[:15].plot(kind='pie',figsize=(10,10))

<AxesSubplot:ylabel='production_companies'>

I investigated the production company that produces the most movies. I did that by plotting a pie char of the top 10 production companies with their counts. I found out that Walt Disney Pictures produced the most movies.

Research Question 5 (Which is the most common tagline?)

movies_df['tagline'].value_counts()[:10].plot(kind='pie',figsize=(10,10))

<AxesSubplot:ylabel='tagline'>

# Length of movies with budget
plt.hist(movies_df.budget[movies_df.runtime > 100], 25, facecolor="b", alpha=1,label="long movies")
plt.hist(movies_df.budget[movies_df.runtime < 100], 25, facecolor="g", alpha=0.75,label="short movies")
plt.legend()
plt.xlabel('Budget')
plt.ylabel('Runtime Hours')
plt.title('Histogam')
plt.grid(True)

Conclusions

In the first section, I analyzed the most common genre of the movies. From the analysis, I can conclude that the top 5 genres were Drama, Comedy, Drama|Romance, Documentary,and Horror|Thriller. In the second secion, the director who directed the most movies was John Carpenter. He was followed by Steven Splelberg, Steven Soderbergh, Ridley Scott, and Robert Zemeckis. These were the most top 5 common directors. In the third section, I wanted to determine which keywords were mostly used to search for movies and the keywords used in the movies. The top 5 keywords used were Woman director, duringcreditsstringer, independent film, biography, and christmas. Further, it wa important to determine which different compnanies are known to produce and direct movies. The top five production companies are Universal Pictures, DreamWorks Animation, Marvel Studios, Walt Disney Pictures|Pixar Animation Studios, and Walt Disney Pictures. Finally, I was able to determine that the most common tagline in this analysis is Two films One love.

Limitations

In the first section, the type of genres were too many to analyze and come up with a good conclusion from the analysis. Therefore, I only took an analysis of the top 15 most common types of genres. This was the same with the directors and keywords section where I only analyzed the top 15 common directors and keywords respectfully. A pie plot of the production companies also had many results from the analysis with the data points almost the same length. Therefore, I was only able to plot the first 15 production companies to show the relationship.

References.

https://www.kaggle.com/code/danielleho/exploratory-analysis-netflix-shows-and-movies

DEV Community: Charles Maina Mwangi