Netflix Data Cleaning & Exploratory Analysis Pipeline

An end-to-end Python data pipeline built with Pandas to ingest, audit, clean, and analyze raw Netflix title data directly from the web.

Project Overview

Raw real-world data is often messy, unstructured, and incomplete. This repository demonstrates a production-style data cleaning workflow on raw Netflix title data (~7,800 records).

Instead of ad-hoc script execution, the project uses a modular pipeline architecture with reusable Python functions to clean missing data, standardize data types, extract temporal features, parse compound duration fields, and execute parameterized analysis.

---

Key Features & Modular Functions

Data Cleaning & Preprocessing

• handle_missing_values(df): Imputes missing values in high-cardinality categorical text columns (director, cast, country) with 'Unknown' to retain dataset volume while safely dropping incomplete metadata rows (date_added, rating).
• fix_datatypes_and_features(df): Trims whitespace, casts text dates into pandas datetime64[ns] objects, and extracts added_year and added_month for temporal analysis.
• process_duration(df): Parses mixed text fields (e.g., "90 min", "3 Seasons") into distinct numerical columns (duration_min for Movies, seasons for TV Shows).
• clean_netflix_pipeline(df): Master wrapper that ties all preprocessing steps into a single executable pipeline.

Single-Input Analysis Toolkit

• get_country_stats(country_name): Returns volume, title type breakdowns, and top genres for any country.
• get_director_profile(director_name): Queries all titles, ratings, and runtimes for a given director.
• get_yearly_additions(year): Analyzes platform acquisition trends for a given calendar year.
• search_by_genre(genre): Filters titles by genre keyword and calculates average movie runtime metrics.

---

Project Structure

├── netflix_pipeline.ipynb   # Complete Google Colab Notebook with cleaning & analysis runner
├── netflix_cleaned.csv      # Exported production-ready cleaned dataset
└── README.md                # Project documentation

---

Key Insights from Analysis

• Average Movie Runtime: ~99.3 minutes.
• Max TV Series Seasons: 16 seasons.
• Top Production Hubs: The United States and India lead in total titles available.
• Content Peak: 2018 saw a major influx of new additions (~1,684 titles added in a single year).

---

Quick Start / Usage

Prerequisites

• Python 3.8+
• pandas
• numpy

Running the Notebook

1. Clone this repository:

git clone https://github.com/YOUR_USERNAME/YOUR_REPO_NAME.git

2. Open netflix_pipeline.ipynb in Google Colab or Jupyter Notebook.
3. Execute cells sequentially to pull the raw data, run the master pipeline, and test the analysis toolkit!