Netflix Data Analysis

📌 Project Overview

This project performs Exploratory Data Analysis (EDA) on a Netflix titles dataset containing 8,807 movies and TV shows.

The analysis focuses on understanding Netflix's content library through data cleaning, transformation, aggregation, statistical analysis, and visualization using Python.

🎯 Objectives

Understand the structure and characteristics of the Netflix dataset
Handle missing and inconsistent values
Transform and prepare data for analysis
Compare Movies and TV Shows
Analyze content distribution by country, rating, and year
Study movie duration and identify outliers
Perform statistical and correlation analysis
Create visualizations to identify useful patterns and trends

📊 Dataset

The original dataset contains 8,807 records and 12 columns:

show_id
type
title
director
cast
country
date_added
release_year
rating
duration
listed_in
description

The dataset includes both Movies and TV Shows.

🛠️ Technologies Used

Python
Pandas – Data manipulation and analysis
NumPy – Numerical operations
Matplotlib – Data visualization
Seaborn – Statistical visualization
SciPy – Statistical analysis
Scikit-learn – Machine learning utilities and evaluation metrics
Regex – Text and duration processing

🔍 Data Cleaning & Transformation

The project includes:

Handling missing values in columns such as director, cast, country, rating, duration, and date added
Converting date_added into a datetime format
Extracting year and month from the date
Creating year_added and month_added features
Extracting numeric movie duration using regular expressions
Converting duration into a numerical format
Cleaning special characters from titles
Downcasting numerical columns to improve memory usage

📈 Analysis Performed

Content Analysis

The dataset contains:

6,131 Movies
2,676 TV Shows

Movies make up the larger portion of the Netflix content in this dataset.

Country Analysis

The analysis compares Netflix content across different countries and content types.

The United States has the highest number of titles in the dataset, followed by India and the United Kingdom.

Year-wise Analysis

The number of titles added to Netflix was analyzed by year.

The highest number of titles were added in 2019, with 2,016 titles, followed by 2020 and 2021.

Movie Duration Analysis

The project analyzes movie duration using descriptive statistics:

Mean: 99.53 minutes
Median: 98 minutes
Mode: 90 minutes
Standard deviation: 28.37 minutes
Variance: 804.82

The longest movie in the dataset has a duration of 312 minutes.

Director Analysis

The project also identifies the most prolific directors based on the number of titles associated with them.

Statistical Analysis

The notebook includes:

Mean, median, and mode
Variance and standard deviation
Skewness and kurtosis
Correlation analysis
IQR-based outlier detection

The analysis identified 453 movie-duration outliers, representing approximately 7.4% of the movies.

📊 Visualizations

The project uses Matplotlib and Seaborn to visualize patterns in the dataset.

Examples include:

Netflix content type distribution
Movie and TV Show comparisons
Duration analysis
Content trends
Statistical distributions
Other exploratory visualizations

📊 Visualizations

Movies vs TV Shows

Netflix Titles Added by Year

Movie Duration Distribution

Top 10 Countries by Number of Titles

Top 10 Directors by Number of Titles

📁 Project Structure

Netflix-Data-Analysis/
│
├── Netflix_Data_Analysis.ipynb
├── netflix_titles.csv
├── README.md
│
├── netflix_content_type.png
├── netflix_titles_by_year.png
├── netflix_movie_duration.png
├── netflix_top_countries.png
├── netflix_top_directors.png
└── netflix_top_ratings.png

🚀 How to Run

Clone or download this repository.
Make sure Netflix_Data_Analysis.ipynb and netflix_titles.csv are in the same folder.
Open the notebook using Jupyter Notebook, JupyterLab, or Google Colab.
Run the cells from top to bottom.

The notebook loads the dataset using:

df = pd.read_csv("netflix_titles.csv")

💡 Key Takeaways

Netflix's catalog contains significantly more Movies than TV Shows in this dataset.
The United States has the largest number of titles.
2019 had the highest number of titles added to the dataset.
Movie duration is centered around approximately 100 minutes.
Statistical analysis and visualizations help identify distributions, relationships, and unusual values within the dataset.

👩‍💻 Author

Nyasa Desai

Computer Science Student SCET, Surat

⭐ If you found this project useful, feel free to star the repository!