Movies Dataset — Exploratory Data Analysis

A data analysis project that explores patterns and trends in movies released between 1980 and 2020, using a dataset of 7,668 films sourced from IMDb.

---

Dataset Overview

Source: Kaggle — danielgrijalvas/movies File: movies.csv
Rows: 7,668 movies
Time period: 1980 – 2020

Column	Description
name	Title of the movie
rating	MPAA rating (R, PG, PG-13, G, etc.)
genre	Primary genre
year	Year of release
released	Full release date and country
score	IMDb rating (out of 10)
votes	Number of IMDb votes
director	Director's name
writer	Writer's name
star	Lead actor/actress
country	Country of production
budget	Production budget (USD)
gross	Box office gross revenue (USD)
company	Production company
runtime	Runtime in minutes

Genres covered: Action, Adventure, Animation, Biography, Comedy, Crime, Drama, Family, Fantasy, History, Horror, Music, Musical, Mystery, Romance, Sci-Fi, Sport, Thriller, Western

---

What This Project Does

1. Data Cleaning

• Drops rows missing critical fields: score, votes, writer, country, runtime, rating, star
• Fills missing budget and gross with their median values (chosen because both columns are highly right-skewed)
• Fills missing company with "UNKNOWN" to retain the row without guessing the value
• Removes duplicate rows
• Adds a derived profit column (gross - budget)
• Adds a decade column grouped by 10-year periods
• Converts votes and runtime from float to int

2. Analysis & Visualizations

Genre analysis

• Which genre has the highest average IMDb score?
• Which genre earns the most money on average?

Director analysis

• Top 20 directors by average IMDb score, filtering to only those with more than 3 movies (to avoid one-hit wonders skewing the results)

Correlation studies

• Does runtime affect IMDb score? (~0.4 correlation — weak but positive)
• Does bigger budget lead to higher gross revenue? (strong positive correlation)
• Scatter plot with regression line for budget vs gross

Country analysis

• Top 15 countries by number of movies produced (US leads by a large margin)

Decade analysis

• Which decade had the best average IMDb scores?

Best and worst movies

• Biggest flop (lowest profit)
• Most successful movie (highest profit)

---

How to Run

Requirements

pip install numpy pandas matplotlib seaborn

Run

python app.py

Make sure movies.csv is in the same directory as the script.

---

Project Structure

├── app.py   # Main analysis script
├── app.ipynb   # Main analysis script in .ipynb format
├── movies.csv           # Dataset
└── README.md

---

Key Findings

• Budget and gross are strongly correlated — studios that spend more tend to earn more
• Runtime and score have a weak positive correlation (~0.4) — longer films score slightly higher but it's not a strong rule
• United States dominates production volume, followed by United Kingdom and France
• Directors with consistent output across multiple films tend to score more reliably than one-film wonders