YouTube Data Analysis & Success Prediction Project

Performed data cleaning and preprocessing on YouTube video data using Python, Pandas, NumPy, and Regex

Handled missing values, duplicates, invalid video IDs, data type conversion, and feature extraction from publishing dates

Standardized and cleaned text-based features such as video tags, titles, and languages

Conducted exploratory data analysis (EDA) using Matplotlib and Seaborn to identify patterns in views, likes, comments, tags, languages, and channel performance

Built multiple visualizations including bar charts, scatter plots, heatmaps, line charts, pie charts, and word clouds to highlight engagement trends

Engineered new features such as log-transformed engagement metrics, number of tags, day of publication, and yearly trends

Developed machine learning classification models to predict video success and engagement levels using Random Forest and Logistic Regression

Evaluated model performance using accuracy score and classification metrics

Tools & Technologies

Python, Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn, Regex, WordCloud