OTT Streaming Platform Analytics

An end-to-end data analysis project examining content trends, platform performance, and audience behavior across Netflix, Prime Video, and Hotstar — using SQL, Python, and Power BI.

Overview

This project analyzes a dataset of 2,500 streaming titles across three major OTT platforms to uncover patterns in content quality, platform strategy, audience engagement, and genre performance.

The analysis moves through three integrated stages:

1. SQL — business-question-driven queries on a SQLite database
2. Python — exploratory data analysis, cleaning, and visualization
3. Power BI — an interactive dashboard for stakeholder-facing exploration

Each stage builds on the last — cleaned data from Python feeds into advanced SQL analysis, which in turn feeds the Power BI data model.

Key Findings

• Quality and popularity are nearly uncorrelated (r = 0.18) — rating appears to set a popularity ceiling rather than drive it directly
• Prime Video leads in content volume and engagement, but Netflix converts engagement into popularity most efficiently
• Genre vote share is almost perfectly even across all 7 genres — no single genre dominates audience attention
• Content quality dipped sharply in 2022-2023 before a strong recovery in 2024
• Rating and weighted rating are almost entirely decoupled from the engagement/popularity/trending metric cluster

Full analysis and methodology in FINDINGS.md

Tools & Technologies

• Database: SQLite
• Languages: Python (pandas, matplotlib, seaborn), SQL
• Visualization: Power BI Desktop
• Dataset: 2,500 streaming titles, 19 attributes (ratings, votes, engagement, popularity, trending scores, genre, platform, etc.)

Project Structure

ott-analytics-project/
├── data/              # Raw and cleaned datasets, summary tables
├── sql/               # Business queries and advanced SQL analysis
├── python/            # EDA, cleaning, and visualization scripts
├── charts/            # Exported visualizations (8 charts)
├── powerbi/           # Interactive dashboard (.pbix)
├── README.md
└── FINDINGS.md        # Full write-up of insights and methodology

How to Run

1. Clone this repo
2. Run python sql/01_business_queries.py to build the SQLite database and execute business queries
3. Run python python/eda_and_cleaning.py to generate cleaned data and charts
4. Run python sql/02_advanced_analysis.py for window-function analysis and Power BI summary table exports
5. Open powerbi/ott_dashboard.pbix in Power BI Desktop to explore the interactive dashboard

Dashboard Preview

Dataset

The dataset (ott_movies_clean_unique.csv) contains 2,500 titles split across Netflix, Prime Video, and Hotstar, covering 7 genres, 6 countries, and 5 languages, with metrics for rating, votes, weighted rating, engagement score, popularity score, and trending score.