Prime Data Explorer: Data-Driven Insights on Amazon Prime Content

Problem Statement & Business Objective

Companies and streaming platforms often struggle to understand the dynamics of content performance and audience engagement. With an ever-growing library of movies and TV shows, identifying trends, assessing content quality, and optimizing production strategies can be challenging. This project addresses these challenges by leveraging a comprehensive dataset comprising Amazon Prime Titles and Credits to extract actionable insights. Our objectives are to:

• Assess Content Quality: Evaluate ratings and runtime distributions to gauge audience satisfaction.
• Identify Trends: Analyze release years, genre popularity, and production countries to understand market evolution.
• Enhance Strategic Decision-Making: Provide data-driven insights that help optimize content acquisition, production, and marketing strategies.
• Improve Audience Engagement: Uncover patterns that lead to higher engagement and identify potential areas of negative performance.

Approach & Methodology

The project follows an end-to-end data analysis process:

1. Data Exploration: We begin by loading and exploring the datasets using commands like df.head(), df.info(), and df.describe(include='all') to assess structure, duplicates, and missing values.
2. Data Wrangling & Preprocessing:
   • Cleaning: Duplicate records are removed and missing values are addressed (e.g., filling missing age_certification with 'NR' and imputing missing runtime values using the median).
   • Data Type Conversion: Key fields such as release_year are converted to appropriate data types for accurate analysis.
   • List Processing: String representations of list-like columns (e.g., genres and production_countries) are converted into actual Python lists.
   • Feature Engineering: A new column num_genres is created to capture the number of genres per title.
   • Merging Data: The Titles and Credits datasets are merged using the id column, creating a unified view that links content with its cast and crew details.
3. Visualization & Analysis: Following the Univariate, Bivariate, and Multivariate (UBM) framework, we develop over 20 visualizations that include:
   • Univariate Analysis: Bar charts, histograms, and word clouds to examine individual variable distributions.
   • Bivariate Analysis: Scatter plots, line charts, and box plots to uncover relationships between variables such as ratings and runtime.
   • Multivariate Analysis: Correlation heatmaps, pair plots, and bubble charts to explore deeper interactions and clusters within the data. These visualizations provide insights such as:
   • Trends over release years.
   • Popularity of genres and production regions.
   • Relationships between IMDb and TMDB ratings.
   • Identification of potential areas of underperformance or negative growth.

Dependencies

This project requires the following Python packages:

• Python 3.x
• pandas
• numpy
• matplotlib
• seaborn
• wordcloud
• jupyter

Install all dependencies using:

pip install -r requirements.txt