Omrilevi123/youtube-stats-analysis

Analysis of Global YouTube Statistics 2023 📊

Project Overview

This project explores a dataset of the top 1000+ YouTube channels in 2023. The goal is to identify key factors that contribute to a channel's success, focusing on categories, geographical distribution, and the relationship between views and earnings.

Data Cleaning & Decisions (EDA)

During the Exploratory Data Analysis phase, several key decisions were made:

• Column Cleaning: Stripped whitespaces from column names for technical consistency.
• Handling Missing Values: Rows missing critical information like category or Country were removed. Numerical missing values were filled using the Median to prevent bias from outliers.
• Outlier Detection: Used Boxplots to identify "Mega-Channels" (e.g., MrBeast). These were kept in the dataset as they represent a realistic part of the YouTube ecosystem.

Research Questions & Insights

1. Which content categories are the most popular?

• Bar Chart of average subscribers per category.
• Insight: Categories like [shows] tend to have the highest average subscriber counts, indicating high viral potential.

2. Geographical Distribution of Top Channels

• Bar Chart of the top 10 countries.
• Insight: The USA and India dominate the global YouTube landscape, likely due to their massive population and established digital advertising markets.

3. Correlation between Views and Earnings

• Visualization: Scatter Plot (Log Scale) of Video Views vs. Earnings.
• Insight: There is a strong positive correlation between views and revenue. However, the variance suggests that factors like target audience location (CPM) also play a significant role.

Google colab

Click here to view the Python Notebook

📽️ Project Presentation

Watch the full walkthrough of the project, including the code execution, data cleaning process, and key findings:

Author: Omri Levi
Course: Introduction to Data Science, Reichman University